Naval ISR Fusion Principles,

Operations, and Technologies

 For a complete listing of titles in the
Artech House Intelligence and Information Operations Library,
turn to the back of this book.
Naval ISR Fusion Principles,
Operations, and Technologies

Jim Scrofani
Will Williamson
Jihane Mimih
Ed Waltz
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the U.S. Library of Congress.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.

Cover design by Andy Meaden Creative

ISBN 13: 978-1-63081-894-4

© 2023 ARTECH HOUSE

685 Canton Street
Norwood, MA 02062

All rights reserved. Printed and bound in the United States of America. No part of this book
may be reproduced or utilized in any form or by any means, electronic or mechanical, including
photocopying, recording, or by any information storage and retrieval system, without permission
in writing from the publisher.
All terms mentioned in this book that are known to be trademarks or service marks have been
appropriately capitalized. Artech House cannot attest to the accuracy of this information. Use of
a term in this book should not be regarded as affecting the validity of any trademark or service
mark.
The views expressed in this publication are our own and do not imply endorsement by the
Office of the Director of National Intelligence, the Department of Defense, or any other U.S.
government agency.

10 9 8 7 6 5 4 3 2 1
 We dedicate this book to the men and women of the United States Navy; we
admire their foresight and leadership in developing, testing, evaluating, and
deploying new technologies to expand the horizons of intelligence, surveillance, and
reconnaissance. We wish them fair winds and following seas
Contents

Foreword xi

Preface xv

1 The Naval Intelligence Reconnaissance and

Surveillance Mission 1

1.1 The Domain of Naval Operations 2

1.2 Naval Mission Concepts and ISR Roles 4

1.3 Principles of Distribution, Automation, and Speed 9

1.4 MDA and ISR Operations in Conflict and Warfare 12

1.5 About This Book 17

Endnotes 20

2 Principles of Operations from Seabed to Space 23

2.1 Maritime Awareness 25

2.2 Naval Subsurface and Seabed Domains 31

2.3 Naval Surface and Airborne ISR 34

2.4 Naval Space 35

vii
viii Naval ISR Fusion Principles, Operations, and Technologies

2.5 Naval Cyber 40

2.6 A Seabed-to-Space Scenario 42

Endnotes 51

3 Distributed Maritime Operations 55

3.1 Introduction to DMO 56

3.2 DMO Tactical Considerations 58

3.3 DMO Architecture and Elements 60

3.4 All-Domain C2 Battle Management 62

3.5 A DMO Scenario 64

Endnotes 67

4 Naval Information Fusion Systems 69

4.1 Enterprise-Level Fusion 70

4.2 Information Systems Fusion 75

4.3 Naval C4ISR Challenges 82

Endnotes 84

5 All-Domain Fusion and Operation Challenges 87

5.1 Challenge 1: Spatial Distribution, Association,

and Latency 90

5.2 Challenge 2: Temporal Sample Rate and Dynamic Targets 93

5.3 Challenge 3: Accuracy for Fire Control and Missile

Engagement 98

5.4 Challenge 4: Integrating Cyber Capabilities 103

Endnotes 105

6 Maritime MultiINT Fusion Processes 107

6.1 The JDL Model for Organizing Naval ISR Fusion 109
 Contents ix

6.2 Maritime Object and Situation Assessment

Levels 1 and 2 114

6.3 Maritime Impact or Threat Assessment Level 3 120

6.4 Maritime Distributed Resource Allocation and

Orchestration 122

6.5 Conclusion 126

Endnotes 127

7 Sensor Distribution and Adaptation 131

7.1 Sensor Networks and Grids 133

7.2 Advanced Wireless Networks 137

7.3 Functional Nodes on the Maritime Network 139

7.4 Distributed Data Fusion in Network Operations 144

Endnotes 152

8 The Role of AI, Automation, and Autonomy 155

8.1 Automating Naval Systems 157

8.2 Sensemaking in a Naval ISR Context 161

8.3 C2 Automation 167

8.4 Automation in the All-Domain ISR Fusion Scenario 172

Endnotes 172

9 Distributed Space Maritime Surveillance 177

9.1 Smallsat Constellations for Sensing and Communication 179

9.1.1 Visible Imagery: Planet Labs Flock 180
9.1.2 SAR: Capella 181
9.1.3 RF Signals Collection: HawkEye 360 182

9.2 A Networked Constellation for Ocean Surveillance

and ISR 183

9.3 Contributions to Naval ISR 185

x Naval ISR Fusion Principles, Operations, and Technologies

9.4 Surveillance Constellation Performance with Examples 187

9.5 Performance Study Results 191

9.5.1 Planet FLOCK–Based Solutions 191
9.5.2 Capella-Based SAR Solutions 192
9.5.3 Hybrid MultiINT Design 192

9.6 Conclusions 193

Endnotes 195

10 Future Technologies to Enable All-Domain 199

10.1 Naval Technology Scanning to Avoid Surprise 201

10.2 Undersea and Seabed Surveillance 205

10.3 Quantum Technologies in Naval ISR 213

10.3.1 Quantum Computing 214
10.3.2 Quantum Sensing 216
10.3.3 Quantum Communication 219
Endnotes 221

About the Authors 225

Index 229
Foreword
As professional naval intelligence officers with keen interests in emergent dy-
namic intelligence, surveillance, and reconnaissance (ISR) processes, we needed
to develop a significant appreciation for the challenge of integrating informa-
tion from many sources, and the potential role of technology to contribute
automation to achieve speed and accuracy. Working with a team of engineers in
the 1980s, what came to be known as the Joint Directors of DoD Labs (JDL)
Data Fusion model was developed—a reference architecture that remains today
as a standard for integrating information in intelligence applications. Through-
out the 1980s and 1990s, the model guided the development of many naval
data fusion capabilities as new sensor and digital links enabled even more in-
formation to be fused to provide situation awareness to combat information
centers across operating forces. Our requirement in those days was to integrate
and derive threat information from the seabed to space.
By the early 2000s, it was known that the U.S. Navy had a sound vision
in Network Centric Operations (NCO) and an implementation concept in
FORCEnet, but to achieve enterprise transformation, we needed to solve many
challenges. The first was acquisition reform and transformation to enable rapid
change at large scale—a continuing problem and issue in the DoD to this day.
We needed a level of system-of-system (SoS) enterprise level systems engineer-
ing that was significantly more expansive than prior DoD projects—the closest
was the air traffic network developed by the Federal Aviation Administration
(FAA). We needed the highest level of continuous industry-government col-
laboration at both the enterprise and system/program level to accomplish net-
working and integration at the naval force scale. So, we had to apply software

xi
xii Naval ISR Fusion Principles, Operations, and Technologies

technologies to build composable enterprise services that worked at a large scale

and across the naval networked hubs across the fleet, and many ashore facilities.
Throughout the early years we also spent a lot of time thinking about what the
naval officer needs to do and how it is done with automation—the role today
we call human-machine teaming. This is an area that remains a challenge, how
do humans and machines now coordinate all-domain unmanned and manned
autonomous systems and sensors in what was to ultimately become even more
challenging weaponized ISR?
We recognized early on that the fusion process needed to be more distrib-
uted and adaptive—able to work across the vast network of navy sensors and
sources, serving more distributed command and control (C2) capabilities. We
believed that such a system would provide vastly improved situation awareness
but also needed a means to characterize what and how much we don’t know—
and figure out where we should be searching. The data fusion community was
given many challenges.
Now, two decades after the FORCEnet vison we have made significant
technical and operational progress. Modern navies face challenges by adversar-
ies that seek to control areas of the seas and littorals. Distributed sensors, ubiq-
uitous cyber, and precise, hypersonic-fast long-range weapons enable large-area
denial capabilities. To counter these threats, distributed maritime operations
(DMO) have become more the order of the day that relies even more critically
on secure and assured networks for ISR and BMC2. Naval strategies, opera-
tions, and tactics are supported by complex technologies that are intended to
bring networked centric warfare to the forefront—and continuously transform-
ing naval ISR. This book is about that transformation.
A team of professors at the Naval Postgraduate School that understand
this area well have compiled this comprehensive book to organize and explain
the advanced methods and capabilities in naval ISR, and the impact on naval
operations. The authors are, appropriately, from the school’s Center for Multi-
ple Intelligence (MultiINT) interdisciplinary studies, and they bring a breadth
of knowledge about the various INTs and the processes to bring them together
into a coherent picture of the naval battlespace. In this book, the authors move
from the need for distributed maritime operations, to the enabling ISR tech-
nologies—from satellite constellations to integrated unmanned systems and
quantum sensing. An important feature of this book is the breadth and depth
of scholarly literature reviewed to allow the reader to go beyond the pages of
the book and explore the indispensable technical literature in the field of data
fusion. The authors explain the fundamentals of ISR, the networking of sensors
from seabed to space, and the methods of data fusion—corelating and combin-
ing data from the vast collection network into a correlating view.
 Foreword xiii

This is an important book for the naval officer, developer, or system im-
plementer in advanced ISR, and for those seeking a deeper understanding of the
challenges, operations, and methods of advanced naval ISR systems.

Admiral William O. Studeman, USN, Ret.

March 2023
Preface
It is a time of rapid change in maritime strategy, naval operations, and fleet
tactics. Technology is a key driver for this change as intelligence, surveillance,
reconnaissance, and targeting (ISRT) is enabling faster, more precise, and lon-
ger-range maritime operations. Maritime information superiority is the goal
sought by modern navies to enable this new speed, precision, and reach for the
ocean-going fleet.
In this book we introduce the elements, both ISRT processes and technol-
ogies, that are enabling increasing degrees of maritime superiority and the op-
erations that are enabled by it. We provide comprehensive citations to the open
literature and relevant books to guide the readers who desire to dive deeper into
the wide range of technical subjects that we cover.
We necessarily emphasize the advanced technologies that are enabling
new fleet operations and tactics that include more widely distributed vessels,
unmanned vehicles and vessels, and the distribution of sensors from the sea-
bed to space. While we describe current ISR operations and technologies, we
also emphasize future capabilities that will enable the distributed fleet; many of
these capabilities are not operational at the time of this writing, but we antici-
pate will soon join international navies. Among the advanced topics we cover is
the operational concept of DMO and how advanced ISRT capabilities enable
DMO and its distributed lethality. The U.S. Navy has adopted the term seabed
to space to encompass the notion of sensors and operations across the full range
of domains in which the Navy operates—this book also addresses that breadth
of distribution, as well as the concept of network centric operations.

xv
xvi Naval ISR Fusion Principles, Operations, and Technologies

As a team of coauthors, we are also colleagues in the Center for MultiINT

Studies (CMIS), an interdisciplinary research center in the Electrical and Com-
puter Engineering (ECE) department at the Naval Postgraduate School (NPS),
Monterey, California. We conduct cutting-edge research and produce research
outcomes to advance the capabilities to integrate data and information from
multiple intelligence domains (INTs). This book introduces what we do—from
the seabed to space.
We express our gratitude for the support in this effort from Dr. Douglas
Fouts, chair of the electrical and computer engineering at NPS, our colleagues
at NPS, and to our research sponsors that have always encouraged our efforts in
advancing the state of the art in ISR. We thank Mr. Mark Owen of the Naval
Information Warfare Center–Pacific for his encouragement and guidance in
our research, as well as the encouragement of numerous DoD space experts
that we cannot name here. We are also grateful for the excellent support from
our Artech House editors, and the excellent technical reviewer that significantly
enhanced the manuscript with a comprehensive technical review.
We are honored that Admiral William Studeman, the recognized Navy
pioneer in data fusion, has graciously written the foreword to this book.
1
The Naval Intelligence Reconnaissance
and Surveillance Mission
The navies of the world play a unique role in national security, providing border
security, assured maritime access, and forward deployed deterrence. Navies are
employed to project influence and power, and the maritime domain is viewed
as a maneuver space for that projection. Navies are unique in that they move
and maneuver forces within international waters without prior diplomatic
agreement, and can be self-deploying, self-sustaining, sea-based, and operate
as an expeditionary force. The goal of these forces is to achieve maritime supe-
riority—that degree of dominance of one force over another that permits the
conduct of maritime operations by the former and its related land, maritime,
and air forces at a given time and place without prohibitive interference by the
opposing force Department of Defense (DoD).
A key element of all naval operations is the understanding of the maritime
environment, provided by operations and activities of intelligence, surveillance,
and reconnaissance (ISR), highlighted in this book:

Intelligence, surveillance, and reconnaissance (ISR)—1. An integrated

operation and intelligence activity that synchronizes and integrates the
planning and operation of sensors, assets, and processing, exploitation, and
dissemination systems in direct support of current and future operations.
2. The organizations or assets conducting such activities (DoD JP 2-01)
[1].

The U.S. Navy also emphasizes the close relation of ISR to command
and control (C2), and computers (C4) that direct operations by applying the

1
2 Naval ISR Fusion Principles, Operations, and Technologies

term C4ISR to complete the relationship between the information collection

understanding role (intelligence) and the warfighting response role applying
command, control, communications, and computation (C4).
This chapter introduces the naval mission and the drivers that are placing
demands on new modes of operations and the advanced ISR and data fusion
capabilities that are required to support them, especially in the attributes of ISR
distribution, automation, and speed.

1.1 The Domain of Naval Operations

C4ISR is conducted across the entire naval areas of operations defined by termi-
nology that distinguishes the unique areas that make up the maritime domain,
comprised of oceans, seas, bays, estuaries, islands, coastal areas, and the airspace
above these, including the littorals (Table 1.1). Navies uniquely operate at the
boundary of land, sea, and air and now include the boundaries to space above
and cyberspace. The physical attributes of each of these areas constrain and
influence naval operations and the ISR process.
The domain of maritime operations is also defined in international law
by the United Nations Convention on the Law of the Sea (Figure 1.1) to distin-
guish nationally sovereign areas of the airspace and territorial sea (12 miles from
baseline coastal boundary) out to the economic (activity) exclusion zone out to
200 miles from the coast baseline.
Navies face a difficult maritime target challenge to maintain continual
custody of maritime threats (subsurface to space adversary weapon systems and
naval sensing systems that provide surveillance-to-targeting custody informa-
tion for commanders). This can only be achieved by dynamically orchestrating

Table 1.1
Layers and Waters of the Maritime Domain
The Maritime Domain
Cyberspace Domain Within the Global Information Environment
Space Communication, Navigation, and Intelligence Collection
The Airspace Above
Brown The Littorals Green Blue
Water Landward Seaward Water Water
Navigable Area inland from Area from the Area of coastal The open ocean
rivers and their the shore that can open ocean to the waters, ports, and and the deep
estuaries be supported and shore that must harbors seabed below
defended directly be controlled
from the sea to support
operations ashore
The shallow water floor, waves, currents, vegetation The seabed below
 The Naval Intelligence Reconnaissance and Surveillance Mission 3

Figure 1.1 Maritime zones recognized by international law [2].

the Navy seabed-to-space sensor network to assign sensors to targets optimally

and dynamically. In later chapters we describe this process, referred to as level
4 fusion, in detail. These ISR capabilities are required to have the attributes of
speed, flexibility, agility, and scalability to address the increasing scope of the
maritime environment.
New ISR technology is enabling ubiquitous sensing of the world’s oceans
from all of the domains that go from seabed to space. Platform technologies
coupled with artificial intelligence (AI) to increase autonomy and speed of op-
eration are enabling new classes of sensing systems that will significantly expose
the oceans. Consider the six key domains of this exposure:

• Seabed: The deep ocean (6,000–36,000 ft) is the newest potential do-
main of warfare where undersea cables, undersea infrastructure (oil and
gas), mines, and future deeply hidden weapon systems reside and pro-
vide targets in conflict. This is the realm of deep submersible remotely
operated and autonomous systems.
• Undersea: Traditional undersea warfare depths to 6,000 ft are the oper-
ating areas of manned and unmanned submersible weapons and sens-
ing systems; this is the realm of submarine and antisubmarine warfare
(ASW).
• Surface: Surface ships and unattended sensors (buoys, autonomous sta-
tion–keeping sensor nets, etc.) provide surface level, overt, and covert
maritime sensing.
• Information and cyber: This nonphysical domain has a presence across all
of the physical domains. The U.S. DoD defines the information envi-
ronment (IE) as the aggregate of individuals, organizations, and systems
that collect, process, disseminate, or act on information, a heterogeneous
4 Naval ISR Fusion Principles, Operations, and Technologies

global environment where humans and automated systems observe, ori-

ent, decide, and act on data, information, and knowledge. DoD then
defines cyberspace as a global domain within the IE consisting of the
interdependent network of information technology, infrastructures, and
resident data, including the internet, telecommunications networks,
computer systems, and embedded processors and controllers. Cyber-
space operations targeting the physical and informational dimensions of
the IE also have an impact in the human cognitive dimension [3].
• Air: Manned and unmanned airborne platforms are tasked to cover key
areas of naval operafor patrol observation and influence.
• Space: Individual spacecraft, constellations, and clusters of networked
spacecraft provide persistent coverage of the world’s oceans (primarily
radar and electro-optical imaging, and signals collection).

1.2 Naval Mission Concepts and ISR Roles

Navies uniquely operate on the world’s seas and enable operations beyond a na-
tion’s sovereign territory, extending defense and projecting a nation’s power to
defend and influence in the nation’s vital interests. This also engages the naval
force with foreign nations in cooperation and partnerships at sea (e.g., protect-
ing seaways, interdicting illicit operations), and ashore (defense partnerships,
training, etc.).
Navies must be prepared to operate day-to-day and escalate/de-escalate
across the spectrum from competition to lethal combat (Figure 1.2).
A number of key operational concepts and the terms associated with them
must be understood to appreciate the drivers and demands that are being placed
on future naval ISR systems:

• Maritime superiority is that degree of dominance of one force over an-

other that permits the conduct of maritime operations by the former

Figure 1.2 The competition-to-conflict spectrum [4].

 The Naval Intelligence Reconnaissance and Surveillance Mission 5

and its related land, sea, and air forces at a given time and place without
prohibitive interference by the opposing force.
• Maritime domain awareness is the effective understanding of anything
associated with the maritime domain that could impact the security,
safety, economy, or environment [5].
• Information dominance is the operational advantage gained from fully in-
tegrating the Navy’s information functions, capabilities, and resources to
optimize decision-making and maximize warfighting effects. It provides
the ability to seize and control the information domain high ground
when, where, and however required for decisive competitive advantage
across the range of Navy missions [6].
• Sea control is the employment of naval forces, supported by land and
air forces as appropriate, in order to achieve military objectives in vital
sea areas. Such operations include destruction of enemy naval forces,
suppression of enemy sea commerce, protection of vital sea lanes, and
establishment of local military superiority in areas of naval operations.
See also land control operations (U.S. DoD) [7].
• Distributed maritime operations (DMO) is an operational concept that
enables widely dispersed naval units to perform sensing, command, and
control, and weapon activities such that the distributed platforms act as
a coherent whole. This is related to:
• Distributed lethality is an operational concept to more widely distrib-
ute lethal weapon systems across surface ships (“if it floats, it fights”)
allowing all surface units to contribute to coordinated and indepen-
dent surface strike actions.
• Dynamic force employment (DFE) is an operational concept that de-
ploys naval forces over wider and more diverse sets of environments,
rather than establishing regular maritime patterns of operation.
• All-domain ISR is an ISR that encompasses and integrates information
from all domains of the maritime environment; sensors and sources
from seabed to space to provide commanders with the most complete
picture of adversary activities.

In addition to these operational concepts, maritime operations are con-

fronted by significant threat challenges that are the basis for these new concepts
of operation:

• Great power competition is a geopolitical theory that characterizes the

world as a competition that is taking place not only in political, military,
6 Naval ISR Fusion Principles, Operations, and Technologies

economic, social, and information domains, but also in the domain of

ideas (or ideologies) about the state of the world, the goals for the world,
and what should be done to achieve those goals. It is a competition for
influence, and therefore power [8].
• Gray zone competitive strategies is the gray zone between peace and com-
bat and includes competitive strategies that seek to gradually alter the
status quo of power by employing unconventional elements of state
power to influence and assert control. Gray zone operations are not con-
ducted directly by the military, but by proxies (unidentified groups, dis-
guised military, law enforcement, civilian organizations, sympathizers)
using the pretexts (e.g., righting wrongs, protecting certain innocents).
Academics have specifically identified the adoption of the gray zone op-
erational approaches in great power competitions to mitigate threats to
contested maritime areas.

Russia and China are increasingly turning to these tactics to pursue ob-
jectives on their periphery without reaching the level of violence that could
provoke a United States or allied response. In gray zone operations, an aggressor
denies its adversary physical access to contested areas at sea or ashore using ci-
vilian or paramilitary forces, spreads disinformation via social media to foment
protests and insurgencies, and degrades the defender’s computer networks and
sensors through cyber and electronic warfare attacks. For China, these gray
zone tactics represent an attempt to pursue its interests in the near-term with-
out drawing a full-scale response by the United States, its allies, and others. In
contrast, gray zone approaches are a long-standing component of Soviet and
Russian military and intelligence operations [9].

• Antiaccess/area-denial (A2/AD) is an operating environment protected

by: (1) antiaccess enemy actions that inhibit military movement into
a theater of operations, and (2) area-denial activities that seek to deny
freedom of action within areas under the enemy’s control [10].
• Hypersonic threats are weapon systems (missiles) that travel at speeds of
more than five times the speed of sound (Mach 5, approximately 3,800
mph) and therefore reduce their exposure time to detection, tracking,
and intercept. The primary categories include [11]:
• Hypersonic glide vehicles (HGV) are launched from a rocket before
gliding to a target;
• Hypersonic cruise missiles are powered by high-speed, air-breathing
engines, or scramjets, after acquiring their target.
 The Naval Intelligence Reconnaissance and Surveillance Mission 7

• Swarm threats and tactics is the employment of weapons in coordinated

attacks with unmanned weapon systems that saturate defenses to over-
whelm sensing and defensive weapons; this complex integrated threat
drives the need for integrated air and missile defense to deal with smaller
signatures, mass raids, shortened response timelines, and longer ranges.

These environmental and operational factors have driven the development

of goals, objectives, and implementing strategies and principles for naval opera-
tions and ISR. The goal-to-implementation structure (Figure 1.3) for the U.S.
Navy has been articulated in a series of documents that begin with two goals
[12]. Maritime dominance is the operational goal that requires the achievement
of information superiority. Superiority in the information domain drives the
need for operational maritime domain awareness (MDA) which requires all-
domain ISR. Figure 1.3 illustrates the key principles in the goals, operational
objectives, and implementing MDA and ISR approach. The flow-down from
an abstract goal (superiority of information) to practical ISR implementations
is based on several assumptions: (1) better information (speed, accuracy, cover-
age) yields improved situation awareness, (2) this leads to improved decision-
making, and (3) this results in greater lethality and impact.
Maritime dominance requires the maritime force to achieve fundamen-
tal operational objectives in defense, offense, and battlespace awareness pro-
vided by all-domain ISR and the components of meteorology, oceanography,

Figure 1.3 Goals-to-ISR implementation flow-down.

8 Naval ISR Fusion Principles, Operations, and Technologies

intelligence, cryptology, communications, networks, space, and electronic war-

fare (Table 1.2).
Dominance requires the precise synchronization of these defensive, offen-
sive, and ISR activities in time and across domains. For example, consider how
a maritime operation may require ISR capabilities to:

1. Conduct persistent surveillance of a large ocean area by orchestrating

a series of small satellites (smallsat) to locate maritime combatants.
Concurrently coordinate unmanned air reconnaissance missions to
locate dark vessels;
2. Monitor the electromagnetic spectrum for maritime radar and beacon
emissions, communications, and monitor or penetrate cyber channels.
Exploit information extracted from these channels;
3. Correlate and associate the information obtained across the space, cy-
ber, electromagnetic, and physical domains to estimate the maritime
situation (surface and subsurface vessels, aircraft, etc.), activities, and
events;
4. Identify information gaps and potential threats. Issue collection tasks
to orchestrate additional collection;
5. Place the information in the context of weather, the subsurface terrain
and bathymetry, and the current geopolitical and maritime situation;
6. Estimate the current situation based on current observations and prior
knowledge of the capabilities, patterns of operations, and intent of
adversaries;
7. Develop and disseminate to commanders the timely and accurate
common operating picture of Blue and Red (adversary) objects and
projected behaviors and threats;

Table 1.2
Key Operational Objectives of Maritime Superiority
Aspect Operational Objectives
Awareness by ·· Assure knowledge of status of distributed Blue forces;
all-domain ISR ·· Reduce uncertainty, locate, and track maritime objects (Red and neutral);
·· Know the maritime situation and predict feasible next situations;
·· Enable timely, accurate, and decisive command decisions.
Defense ·· Protect, defend, and recover own IE;
·· Increase the enemy’s friction by creating uncertainty.
Offense ·· Degrade, disrupt, deceive, deny, and exploit enemy’s IE to introduce fog;
·· Mass distributed forces decisively engage the enemy’s critical vulnerabilities
to degrade and remove the enemy’s center of gravity.
 The Naval Intelligence Reconnaissance and Surveillance Mission 9

8. Support commander decision-making to manage ISR collection as

well as to plan the application of force using both nonkinetic and ki-
netic actions, and support C2 to plan, direct, monitor, and assess the
employment resources.

Navies perform many different kinds of operations and these influence

the demands placed on ISR capabilities that are required to support opera-
tions. Representative naval operations and the demands on ISR systems are
summarized in Table 1.3 to highlight the wide range of capabilities anticipated
for surveillance and reconnaissance across a wide range of domains. Maritime
operations are not performed alone, and navies often operate as an element of
a joint force with air, space, and ground elements of other national or partner
forces [13].
In joint operations, navies distinguish between organic ISR resources
(those owned by the local force), networked ISR resources that are shared across
a network of users, and national ISR assets that are elements of the national
technical means (NTM) of intelligence collection [14]. The U.S. Navy specifi-
cally develops methods for tactical exploitation of national capabilities (TEN-
CAP) to support tactical naval forces. TENCAP projects focus on maritime
information superiority by [15]:

• Improving procedures by which national systems products are made

available to tactical forces;
• Developing new methods for processing national system data into tacti-
cally useful information;
• Formulating new concepts for using existing national systems in sup-
port of tactical operations and conducting tests and demonstrations to
assess the effectiveness of these concepts;
• Prepare tactical impact statements to influence the design of future na-
tional systems to assure the needs of tactical users receive appropriate
priority.

1.3 Principles of Distribution, Automation, and Speed

To achieve maritime information superiority on a global scale, ISR requires a
wide distribution of sensing capabilities, automation of the sensing-collection
process to provide speed from collection to response. In this section we intro-
duce the principles of these three characteristics required for advanced ISR.
10 Naval ISR Fusion Principles, Operations, and Technologies

Table 1.3
ISR Demands Across Representative Maritime Operations
Operations Operational Role ISR Role
Surface Plan and direct surveillance of the ·· Conduct surface surveillance from
warfare maritime domain, interdict and conduct space, air, or surface sensing.
(SUW) strikes by aircraft and missiles to ·· Perform over-the-horizon target
destroy or neutralize enemy naval acquisition, tracking, and targeting.
surface forces and merchant vessels.
Air and Counter both air and missile threats ·· Perform air surveillance in moving
missile by directing a combination of theater missile engagement zone (MEZ);
defense counterair and integrated air and missile ·· Acquire, track, and handoff high-speed
(AMD) defense (IAMD) weapons. targets.
Antisubmarine Find, fix, track, target, and engage ·· Conduct broad-area maritime
warfare enemy submarines; monitor, track, and surveillance.
(ASW) engage enemy submarines in port or ·· Apply persistent, national, and joint
transiting to operating areas as well as air, space (e.g., EO, IR, radar), surface,
conduct active searches in operating and subsurface (e.g., acoustic,
areas. nonacoustic) intelligence collection
systems.
Mine Identify engagement opportunities ·· Apply ISR to detect mining ops,
warfare to employ friendly mining capability, localize threat areas, and plan safe
(MIW) preclude adversaries from effectively transit Q routes.
employing maritime mining, and defeat ·· Apply ISR to guide mine-sweeping
the minefield. operations.
Strike Destroy or neutralize targets ashore, ·· Apply space, air, human, or other
operations including attacks against strategic or sources to localize ashore activities
tactical targets, from which the enemy and targets.
is capable of conducting or supporting ·· Track movements (e.g., convoys) and
air, surface, or subsurface operations processes (e.g., fueling, weapon
against friendly forces. basing) that pose threats.
·· Identify noncombatant activities.
Amphibious Embarkation and debarkation of landing ·· Perform overwatch ISR to identify
operations forces from/to ships to land. threats complex amphibious ops.
Maritime Monitor, query, and board merchant ·· Obtain departure and in-transit
intercept vessels in international waters to intelligence to track and identify
operations enforce sanctions. suspicious vessels.
(MIO)
Maritime Establish conditions for security and ·· Coordinate with foreign intelligence
security protection of sovereignty in the maritime partners to identify activities, routes,
operations domain (e.g., counterpiracy, counter and vessels.
(MSO) maritime-related terrorism, counter ·· Obtain departure and in-transit
proliferation, transnational crime, illegal intelligence to track suspicious
seaborne migration, or trafficking). vessels.
·· Provide intelligence for visit, board,
search, and seizure (VBSS) activities.
Sea-based Establish an at-sea base for ships or ·· Provide supporting ISR for the seabase
operations platforms to provide a scalable and and all operations from the base to
mobile capability to exercise C2 or ashore and afloat activities.
provide strike, power projection, fire ·· Provide intelligence to seabase
support, and logistic capabilities. commanders on foreign responses.
 The Naval Intelligence Reconnaissance and Surveillance Mission 11

Table 1.3 (continued)

Operations Operational Role ISR Role
Counterdrug Support federal, state, and local law ·· Coordinate with law enforcement
operations enforcement efforts to disrupt the intelligence operations, patrol to
transport and/or transfer of illegal drugs detect targets, monitor, and track
into the United States. at-sea traffic.
Foreign Provide rapid national response in ·· Perform overwatch ISR to identify
humanitarian support of humanitarian and natural threats to assistance activities,
assistance disaster relief operations. assets, and personnel.
Riverine Deploy small, armed, patrol craft to ·· Provide supporting intelligence by
operations provide maritime security in inshore, space, air, and land-based unattended
coastal, and riverine areas by patrol, sensor systems.
surveillance, interdiction, and ·· Provide supporting inshore human
destruction of waterborne and land intelligence.
threats.
From: [16].

Distribution
The first attribute describes the structure of advanced MDA ISR systems. As
DMOs move away from concentrated battle groups and MDA drives the need
for extended surveillance coverage, sensors and weapons are distributed across
a wide maritime area. Following the principles of network centric warfare
(NCW), developed in the 1990s, sensors and weapons across all naval platforms
operate on a network to provide an organic response to threats. The network
concept is built on four fundamental principles that are applied to MDA [17]:

1. A robustly networked set of collectors, processors, and analysts im-

proves information sharing;
2. Information sharing and collaboration to enhance the quality of infor-
mation and shared situational awareness;
3. Shared situational awareness that enables self-synchronization across
sensing platforms;
4. These, in turn, dramatically increase effectiveness—particularly the
ability for ISR to respond to dynamic changes in the environment.

NCW is inherently an information superiority concept of operations

with the goal to achieve shared awareness, increased speed of command, higher
tempo of operations, greater lethality, increased survivability, and a degree of
self-synchronization. Sensors are shared so all platforms see a common operat-
ing picture; weapons are also shared so the network can respond dynamically as
threats change [18].
12 Naval ISR Fusion Principles, Operations, and Technologies

Automation
The scope and scale of MDA drives the need for higher automation in the ISR
process. The scope, difficulty, and range of tasks requires higher dimensionality
of sources and often requires MultiINT sources to be associated and combined
to locate and identify targets such as vessels of interest. The global scale of MDA
demands large areas of coverage and numbers of target vessels. These factors
drive that need for increased automation and the orchestration of sensors and
processing of information and deliver timely intelligence.
Automation and autonomy offer two primary benefits: (1) improved safe-
ty, efficiency, and cost, and (2) expanded coverage beyond what human teams
can accomplish, and increased speed, often beyond human performance for
complex tasks. The MDA ISR challenge requires the latter to rapidly assimilate
MultiINT maritime data, track and identify targets, identify threatening activi-
ties, and nominate targets.

Speed
While the first two attributes dealt with the ISR structure, scope, and scale,
this attribute deals with the need to operate at a speed faster than the adver-
sary’s observe-orient-decide-act (OODA) loop—creating a lag in the adversary
response. The distribution of sensors and introduction of a network increases
the latency from sensing to fusion; this effect of distribution to achieve greater
coverage must be overcome to maintain coverage and speed. “How much speed
is required?” is the dominant operational question to implement distributed
MDA and to conduct DMO. Speed is a relative requirement, and it is relative
to operational sensing, network, and weapons performance, as well as the ad-
versary speed of operations. Therefore, latency from sensor-to-decision maker
(or shooter) is a key factor in the success of highly automated MDA systems.

1.4 MDA and ISR Operations in Conflict and Warfare

ISR activities directly contribute to the goal of MDA defined earlier as the ef-
fective understanding of anything associated with the maritime domain that
could impact the security, safety, economy, or environment. The United States
emphasizes the global scope of its interest in the maritime domain, and ac-
knowledges the MDA for the global maritime domain requires cooperation
among security and commerce organizations and across nations [19]:
This extends to the ISR capabilities that require sharing of global mari-
time information and intelligence to cover the world’s seas. The MDA chal-
lenge is significant in scope and scale, to include [19]:

• Persistently monitor the global maritime domain:

 The Naval Intelligence Reconnaissance and Surveillance Mission 13

• Identified areas of interest;

• Vessels and craft;
• Cargo;
• Vessel crews and passengers.
• Maintain data on vessels, vessel tracks, facilities, and infrastructure;
• Collect, fuse, analyze, and disseminate information to decision makers
to facilitate effective understanding (e.g., detect anomalies, trends, and
patterns, explain situations, predict behaviors, and warn of illegal activi-
ties and threats to maritime security);
• Develop and maintain data on MDA-related mission performance.

We further distinguish the MDA objectives of seeking what is observable

and known (situational awareness), as well as what is anticipated or suspected
(threat awareness) [20]. The MDA process has been characterized by a number
of functional models developed to explain the method to collect data, analyze
the data to develop maritime situation awareness, and further understand ac-
tivities that provide threat awareness; we consider two well-known models here
to illustrate the MDA process [21].

The OODA Loop

The OODA model developed by John Boyd provides a closed loop model from
a command-and-control decision-making perspective [22]. Boyd emphasized
the loop as a cycle to be operated at a tempo (speed) that outpaces the adversary.
The stages of the cycle related to the MDA ISR process are:

• Observe is the ISR process by which ISR sensors and sources collect
information about the maritime environment. The observation process
is guided by direction from the act process to collect information to
improve the orientation and decision processes.
• Orientation is the process that places the observed information in con-
text, cross-correlating information, comparing information to mental
models of maritime behavior, and projecting behavior into possible fu-
tures. Here, intuition and deliberate reasoning enable a situation to be
estimated and alternatives to be considered.
• Decide is the process of applying judgment to choose actions from al-
ternative views of the situation, alternative actions, and the range of
outcomes of those choices. Contextual factors (weather, geopolitical,
logistics, etc.) and experience come to play in this process of judgment.
14 Naval ISR Fusion Principles, Operations, and Technologies

• Act performs the operations in the maritime environment that respond

to the current situation: (1) conduct additional surveillance and recon-
naissance, (2) probe the environment to test hypotheses, (3) commu-
nicate with third parties, (4) engage maritime targets, and (5) conduct
follow-up assessments.

The results of actions are observed and the cycle reorients and makes new
decisions. The detailed functions of the manual MDA ISR process are depicted
in the OODA loop in Figure 1.4 to illustrate the use of the model as an organiz-
ing framework to describe the command-and-control functions and relation-
ships of the MDA process. The example MSO activity receives tips about a
suspicious vessel of interest (VOI) and creates a VBSS mission. The flow begins
with observe and orient as a tip is received at the Maritime Operations Center
(MOC) about a specific VOI. The tip information is assessed to determine
credibility and the VOI is located (if possible) in the current and historical track
knowledge base (SEALINK). In the decide phase, the MOC director guides the
decision-making process that involves requests for information (RFIs) to assess
the availability of resources (e.g., ships or aircraft to perform reconnaissance to
locate the vessel, and ships available along the projected track to intercept and
board).
The MOC director develops the course of action (COA) and issues the
orders to conduct reconnaissance, intercept, and board. The act phase carries
out the COA and boards the ship, collecting biometric data from those onboard
and sending the data ashore for crew identification and VOI analysis. Results
can feed back to the decide phase causing additional actions. If the VOI remains
suspect, it may be placed on the watchlist database for subsequent tracking.

The Joint Directors of DoD Laboratories (JDL) Fusion Model

In the late 1970s, the JDL developed a model to describe the fundamental
elements of the intelligence and fusion process, focusing on the core of the
orient and decide elements of the OODA loop. The four-level model as origi-
nally envisioned is a fundamental description of four canonical elements of hu-
man cognition, applied to the intelligence data fusion process, a core element
of ISR. The elements describe the human (manual) cognitive process as levels
or degrees of activities that step through a cycle [24]. As described initially,
it is a fundamental process model that describes human intelligence analysis,
computer-aided analysis, medical diagnosis, forensic analysis, and many other
endeavors—not just computer data processing.
The original JDL model is based on (or, similar to) a fundamental model
of cognition that has the stages (or levels) of sensation, perception of objects,
 The Naval Intelligence Reconnaissance and Surveillance Mission 15

Figure 1.4 Top-level MDA functions for VBSS mission in the OODA loop model [23].

perception of situations of objects over time, and cognition (reasoning) that

constructs an understanding of the situation, followed by decision and action.
The highest-level model distinguishes System 1 cognition (intuition: fast, per-
formed in parallel, automatic and effortless, slow learning, and emotionally
influenced) and System 2 cognition (reasoning: slow, performed in parallel,
controlled and requiring effort, flexible, and neutral) [25].
The basic four JDL levels are:
16 Naval ISR Fusion Principles, Operations, and Technologies

• Level 1, entity assessment: Estimation and prediction of entity parametric

and attributive states (i.e., of entities considered as individuals);
• Level 2, situation assessment: Estimation and prediction of the structures
of parts of reality (i.e., of relations among entities and their implications
for the states of the related entities);
• Level 3, impact assessment: Estimation and prediction of the utility/cost
of signal, entity, or situation states including predicted impacts given a
system’s alternative courses of action;
• Level 4, performance assessment: Estimation and prediction of a system’s
performance as compared to given desired states and measures of effec-
tiveness. This also includes the feedback to manage resources (levels 1–3
processing and sensors/sources in the closed loop).

Figure 1.5 compares a simple cognition model and the closed-loop JDL
model.
The robotics discipline has adopted a similar model, of course, because
this is quite fundamental, and represents the fundamental elements of thinking
and autonomy. The cognitive functions focus sensing and processing to reduce
uncertainty about the situation and threats.
More recent revisits to the model have added new functions: (1) a level
0 (signal/feature assessment) that precedes level 1 and combines raw signals or
features from multiple sources to detect and identify objects and their states, (2)
a level (or function) 5 (user refinement) that is the human machine interface
(HMI), and (3) a function 6 (mission management) that includes the mis-
sion actions beyond sensing and collection [26]. Waltz introduced the comple-
mentary nature of machine learning (or ML, referred to as data mining in the
1990s) and data fusion to illustrate how inductive ML processes discover and

Figure 1.5 A cognitive model compared to the JDL fusion model.

 The Naval Intelligence Reconnaissance and Surveillance Mission 17

refine models of targets for deductive fusion processes to detect those targets in
operational data (Figure 1.6) [27].
The JDL model has provided a basis for distinguishing the processing
stages and implementing algorithms for five decades and remains the standard
for those in the ISR discipline. The JDL model is adopted as the framework
used throughout this book as the common reference for ISR processes.

1.5 About This Book

This text is designed to help naval officers and technical professionals to under-
stand maritime threats and advanced naval ISR to support DMO to counter
complex threats. The book describes the complexity of achieving seabed-to-
space sensor coordination on a distributed network, the fusion of ISR infor-
mation, and decision-making support. We focus on the methods to apply
advanced technologies in fusion: sensing orchestration, distributed data corre-
lation and association, data combination, and distribution. The book is suitable
for naval officers in the ISR, intelligence, space, ASW, electronic warfare (EW),
cyber, and surface warfare disciplines who seek an in-depth understanding of
advanced ISR operations and technologies. It is designed to provide a compre-
hensive background for naval and industry managers and engineers planning
and developing advanced naval systems—subsurface, surface, air, cyber, and
space.

Figure 1.6 Complementary ML and data fusion processes. (From: [28]. Figure reprinted with
permission from IEEE.)
18 Naval ISR Fusion Principles, Operations, and Technologies

The book is organized to introduce the new maritime threats, policies,

and operations that will place demands on future ISR systems before describing
the methods and technologies that will be required to meet future naval threats.
The chapters are organized in a framework (Figure 1.7) that first intro-
duces the mission, role of ISR in Chapter 1, and the challenges in maritime
environments in Chapters 2–5. The emphasis in this first section is in the chal-
lenges introduced by all-domain ISR and the distribution of sensing and opera-
tional assets. The second half of the book, Chapters 6–10, describe the techni-
cal operation of distributed ISR and the necessary technologies to enable highly
automated, distributed sensing across all domains. At appropriate places in the
book, example scenarios are provided to illustrate the operations and challeng-
es. We also provide throughout quantitative measures of operating parameters,
typical performance, and effectiveness of the scenarios provided.
Chapter 2 introduces the demand for ISR to support coordinated naval
operations in all domains from seabed to space. The naval operations in the do-
mains of seabed, subsurface, surface, cyber, air, and space are introduced with a
focus on how they each contribute to MDA. The demands for ISR across these
domains are explained, as well as the challenge of delivering a timely and com-
prehensive awareness to provide commanders with an understanding of an ad-
versary’s coordinated operations. An all-domain operations scenario is provided
to illustrate how all domains can conceptually be integrated to work together.
Chapter 3 explains the recent doctrine of DMO and its relationship to de-
mands on ISR for widely distributed sensing and force elements, such as DMO
tactics, architecture, and elements. The necessity of a communication grid to
connect distributed nodes, node data storage, processing power, and technol-
ogy stacks are explained as well as the need for data and analytic strategies to

Figure 1.7 Organization of the chapters move from mission to ISR fusion technologies.
 The Naval Intelligence Reconnaissance and Surveillance Mission 19

enable rapid and accurate information to enable decisive action. The chapter
concludes with a DMO scenario to illustrate the functions and operations in
the concept.
Chapter 4 provides an overview of the current U.S. Navy ISR fusion ar-
chitecture as a baseline reference and then projects the C2 challenges that drive
the ISR requirements. The chapter provides an introduction to maritime ISR
systems and a baseline architecture for a more centralized sensor and informa-
tion fusion.
Chapter 5 explains the challenges to perform all-domain ISR, focusing on
three challenge areas. The first challenge is reconciling the spatial distribution
of platforms, associating observations and latency across distributed networks.
The second is utilizing sensor observations with low temporal sample rates and
highly dynamic targets (e.g., fleeting, hypersonic, and swarming). The third
challenge is achieving the persistence required for surveillance and the accuracy
required for C2 and fire control.
The second section of the book details the systems and technologies nec-
essary to enable more distributed operations.
Chapter 6 introduces the functional requirements for naval MultiINT
fusion systems and the technologies that implement them. The chapter briefly
introduces the JDL model for organizing naval ISR functions. Then the core
functions of object and situation refinement, threat refinement, and resource
orchestration are explained. The chapter focuses on the necessity for orchestrat-
ing distributed resources to effectively support DMO.
Chapter 7 explains the necessity of adaptive networks to route sensor data
across a grid of nodes to users: other sensors, fusion nodes, and command nodes
that perform operational decision-making. The principles of network operation
and security are detailed for the grid.
Chapter 8 introduces the emerging role of AI, automation, and autonomy
in naval ISR systems. The rationale and elements of automation of the ISR
network are explained, focusing on automation and intelligence in sensing and
sensemaking. An all-domain automated operation scenario illustrates the con-
tributions of automation and autonomy to DMO.
Chapter 9 focuses on the new global maritime reconnaissance and sur-
veillance opportunities provided by distributed space sensing by constellations
of smallsat constellations, particularly the revolutionary changes in maritime
transparency. The chapter illustrates the game-changing contribution of persis-
tent maritime surveillance and its impact on DMO.
Chapter 10 concludes the book with an overview of future technologies
that will enable even more advanced capabilities in all-domain ISR.
20 Naval ISR Fusion Principles, Operations, and Technologies

Endnotes
[1] Joint Publication 2-01, Joint and National Intelligence Support to Military Operations,
U.S. DoD Joint Chiefs of Staff, July 5, 2017.
[2] Law of the Sea: A Policy Primer, “Maritime Zones” The Fletcher School of Law and Di-
plomacy, 2017, Chapter 2, https://sites.tufts.edu/lawofthesea/chapter-two/.
[3] The IE and its subset and cyberspace are defined by the DoD in Joint Publication (JP)
1-02, the Department of Defense Dictionary of Military and Associated Terms. Notice
that the definitions distinguish between three fundamental domains: the physical domain,
an information domain (which includes cyber), and the human cognitive domain. See the
“DoD Strategy for Operations in the Information Environment,” June 2016.
[4] “A Design for Maintaining Maritime Superiority,” Version 2.0, U.S. Navy, December
2018, p. 6.
[5] Defined in “Amendments to the International Aeronautical and Maritime Search and
Rescue (IAMSAR) Manual,” International Maritime Organization, T2-OSS/1.4 MSC.1/
Circ.1367, May 24, 2010.
[6] “Navy Strategy for Achieving Information Dominance, 2013; The U.S. Navy’s Vision for
Information Dominance,” 2010. See also “Information Superiority Vision,” Department
of the Navy, February 2020.
[7] “Surface Force Strategy: Return to Sea Control,” Surface Force Command, 2018.
[8] The concept of great power competition is articulated in the U.S. Department of Defense,
“National Security Strategy of the United States of America,” Washington, D.C., Decem-
ber 2017.
[9] Clark, B., et al., “Restoring American Seapower: A New Fleet Architecture for the United
States Navy,” Center for Strategic and Budgetary Assessments, 2017, p. 12, https://csba-
online.org/uploads/documents/CSBA6292-Fleet_Architecture_Study_REPRINT_web.
pdf.
[10] Krepinevich, A. F., “The Military-Technical Revolution: A Preliminary Assessment,”
Center for Strategic and Budgetary Assessments (CSBA), 2002, p. 1.
[11] Congressional Research Service, Hypersonic Weapons: Background and Issues for
Congress, R45811, updated March 17, 2020.
[12] Above these Navy documents, the United States has Presidential Policy Directive 18,
“Maritime Security,” (August 2012). From this directive derives a National Security Policy
Document (NSPD-41) and a Homeland Security Policy document (HSPD-13) that define
U.S. maritime security policy (open-ocean and homeland coastal-ports, respectively).
These national policy level documents then flow to a National Maritime Security
strategy that guides Navy goals, operational objectives, and implementing strategies. The
Department of Defense Directive 2005.02E, “Maritime Domain Awareness (MDA) in
the Department of Defense,” defines DoD policy and assigns responsibilities.
[13] For example, see the Department of Defense, Joint Publication 3-0: Joint Operations,
Washington, D.C., 2006, p. IV–26.
 The Naval Intelligence Reconnaissance and Surveillance Mission 21

[14] National technical means (NTM) referred to overhead satellite technical collection systems
in the 1970s though they can be applied to other special collection capabilities. See Smith,
C. E., “Watching the Bear: Essays on CIA’s Analysis of the Soviet Union,” Chapter IV, G.
K. Haines, and R. E. Leggett (eds.), “CIA’s Analysis of Soviet Science and Technology,”
CIA Center for the Study of Intelligence, 2007.
[15] Navy Tactical Exploitation of National Capabilities (TENCAP) Unclassified Solicitations
(FY 2015–2020) DISC Working Documents (2020). Naval Digital Integration Support
Cell.
[16] Operations in the table are based on Maritime Operations Joint Pub 3-32, Department of
Defense, June 8, 2018.
[17] Cebrowski, A. K., and John H. Garstka, “Network Centric Warfare: Its Origin and
Future,” Proc. Naval Institute, Vol. 124/1/1,139, January 1998. See the fundamental
literature on NCO at http://www.dodccrp.org.
[18] Garstka, J., F. P. Stein, and D. S. Alberts, “Network Centric Warfare: Developing and
Leveraging Information Superiority,” Washington, D.C.: DOD C41SR Cooperative
Research Program, February 2000, p. 2.
[19] See “National Strategy for Maritime Security,” September 2005, and “National Plan to
Achieve Maritime Domain Awareness for the National Strategy for Maritime Security,”
United States, October 2005.
[20] “Navy Maritime Domain Awareness Concept,” U.S. Navy Chief of Naval Operations,
May 2009.
[21] In addition to the two models described, two other models should be noted: (1) The sense-
hypothesize-option-respond (SHOR) model was introduced in 1981 by Joseph Wohl
(Wohl, F. G., Force Management Decision Requirements for Air Force Tactical Command
& Control, IEEE Transactions in Systems, Man and Cybernetics, SMC11, 9, pp. 618–639
(September 1981)), and (2) the task-collect-process-exploit-disseminate (TCPED) model
that describes the sequence of activities performed in intelligence collection-to-reporting
(TCPED is a technical representation of the more general intelligence cycle model).
[22] Osinga, F., “Science, Strategy and War: The Strategic Theory of John Boyd,” Netherlands:
Eburon Academic Publishers, 2005.
[23] Adapted from the VBSS example provided in Hutchins, S. G., et al. “Enhancing Maritime
Domain Awareness,” Proc. of 13th ICCRTS: C2 for Complex Endeavors, 2008. See also
“Maritime Domain Awareness: Assessment of Current Status,” Proc. of 14th ICCRTS: C2
and Agility, 2009.
[24] “Functional Description of the Data Fusion Process,” Joint Directors of Laboratories,
1991.
[25] Readers familiar with the Kahneman-Tversky cognition models distinguish between two
models, or systems of thought: fast (intuitive) reaction and slow (deliberative analysis).
Kahneman, D., Thinking, Fast and Slow, NY: Farrar, Straus and Giroux, 2011. Kahneman,
D., “A Perspective on Judgment and Choice,” American Psychologist, Vol. 58, No. 9,
September 2003, pp. 697–720.
22 Naval ISR Fusion Principles, Operations, and Technologies

[26] Llinas, J., C. Bowman, G. Rogova, A. Steinberg, E. Waltz, and F. White, “Revisiting the
JDL Data Fusion Model II” International Data Fusion Conference, July 2004. For a more
recent view, see Blasch, E. P., É. Bossé, and D. A. Lambert, High-Level Information Fusion
Management and System Design, Norwood, MA: Artech House, 2012.
[27] Waltz, E. L., “Information Understanding: Integrating Data Fusion and Data Mining
Processes,” Proceedings of the 1998 IEEE International Symposium on Circuits and Systems,
Vol. 6, 1998, pp. 553–556.
2
Principles of Operations from Seabed to
Space
Although the concept of DMO represents a new thrust in today’s naval warf-
ighting scenario, the idea has developed over several decades following techno-
logical advances and historical events. In the 1950s, the first tactical data link
was developed and deployed in the United States in response to the develop-
ment of Soviet antiship cruise missiles. The data link had the advantage of
reducing latency as well as the time to observe, orient, decide, and act, thereby
enabling the warfighters a quick reaction time [1]. In the late 1970s, the Aegis
Combat System, an advanced C2 and weapon control system, was developed
which enabled combat systems onboard a ship to work together to track and
guide weapons. In the 1990s, a cooperative engagement capability (CEC) was
developed to enable raw sensor data to be shared in real time with other combat
units. This sensor data sharing was a transformative step that allowed ships to
respond to a threat based on data detected by another ship’s radar if they were
part of the CEC network. Former Rear Admiral Rodney P. Rempt, director of
theater air and missile defense on the Navy staff, envisioned a future tactical
grid as an agnostic network of weapons and sensors, controlled by a specific
number of nodes, with no restrictions on the location or deployment site of the
weapons, sensors, or controlling nodes [2]. The former chief of naval operations
(CNO), Admiral John M. Richardson emphasizes in the 2016 report “A Design
for Maintaining Maritime Superiority” the need to “adopt a strategy that allows
sea control and maritime superiority to address the threat rising from Russia
and China and also to address ‘blue-water’ scenarios far from land and power
projection ashore in a highly ‘informationalized’ and contested environment”
[3]. Although this event did not directly call for DMO, it has certainly set the

23
24 Naval ISR Fusion Principles, Operations, and Technologies

stage and presented a challenging problem to which DMO is the only solution.
The CNO was confident that even though tough choices were made due to
budget restrictions, innovative and inspiring ideas would emerge as a result of
these choices. Indeed, in 2017, former Vice Admiral Thomas S. Rowden from
the Naval Surface Forces presented his “Surface Force Strategy, Return to Sea
Control” document that discussed the concept of distributed lethality (DL)
that aims at achieving sea control especially in the surface domain. The DL
concept is based on the following pillars:

• Increasing the potency of single warships by networking the firing ca-

pability;
• Geographically spreading out over a wide area the offensive capability;
• Enhancement of combat capability through the allocation of resources
to surface platforms [4].

In 2018, the U.S. Harry S. Truman Carrier Strike Group presented the
concept of DFE whose goal was to “more flexibly use ready forces to shape
proactively the strategic environment while maintaining readiness to respond to
contingencies and ensure long-term warfighting readiness” [5]. The chairman
of the Joint Chiefs of Staff, General Joe Dunford, and Defense Secretary Jim
Mattis pushed implementation of DFE to allow the military to quickly deploy
forces while ensuring that the military infrastructure remained agile and less
predictable [6].
The DMO concept shares the key tenets with DL and carries the prin-
ciples that form the basis of DFE, except that it extends the maritime domain
to encompass all domains including air, subsurface, space, and cyber warfare.
A primary goal of DMO is allowing operational commanders the ability
to treat the entire fleet assets under their control as a single weaponized system.
The individual systems of varying capabilities are integrated to provide both a
unit defense and a collective defense. The integrated package unites diverse and
disparate weapons, technologies, sensors, platforms, and systems into a single
system with capabilities exceeding the sum of the parts. In addition to engaging
and defeating multidimensional threats across all domains in the maritime en-
vironment, the single united weapons system can also conduct offensive strikes
and other fleet engagements [4].
While the U.S. Navy has been open about its conceptual strategies and
operations over the past decades, the needs and technical solutions described
throughout this book are of interest to navies of the world that subscribe to
the conventions of international security on the seas. These nations partici-
pate in organizations and alliances that cooperate, establish standards, and ex-
change thought on maritime security. The interested reader should monitor
 Principles of Operations from Seabed to Space 25

international maritime policy and technology sources on the current views of

international navies, for example:

• The Center for International Maritime Security (CIMSEC) publishes

reader-submitted content specifically on international maritime peace
and security. CIMSEC has 20 international chapters, over 3,000 mem-
bers, subscribers in 60 countries, and has published many contributed
technical articles on aspects of DMO.
• The International Maritime Organization (IMO) provides policy guid-
ance and “machinery for cooperation among Governments in the field
of governmental regulation and practices relating to technical matters of
all kinds affecting shipping engaged in international trade; to encourage
and facilitate the general adoption of the highest practicable standards in
matters concerning maritime safety, efficiency of navigation and preven-
tion and control of marine pollution from ships” [7].
• NATO Allied Maritime Command [8] and the NATO Science and
Technology Organization [9].
• Australian Defence Science and Technology Group reports on maritime
and ISR technologies [10].

2.1 Maritime Awareness

MDA is defined in the National Strategy for Maritime Security as “the effective
understanding of anything associated with maritime domain that could impact
the security, safety, economy, or environment of the United States” [11]. MDA
is a tool that allows navies, coast guards, and other maritime-based organiza-
tions to monitor the oceans, assess threats, and ensure that waters are safe and
free from any criminal activities. The air and ground domains have designated
check points and well-defined infrastructure and support systems as well as
rules and policies. Although developed over centuries, the maritime domain has
always lacked structure and cohesion of operation, making open oceans even
more vulnerable. Just weeks after 9/11, a 50 ft–long boat offloaded over 200 il-
legal immigrants right in downtown Miami, thereby exposing the vulnerability
of the security of the nation as well as putting its economic and transportation
systems at risk [12]. The inability to predict or keep custody of targets of inter-
est before they become a threat along with increasing concern about terror-
ism, drug smuggling, human trafficking, and piracy, resulted in placing MDA
and maritime security high on several nations’ security agendas in the early
2000s [13]. MDA aims at improving ocean transparency by learning ship traf-
fic trends, detecting illegal or nefarious activities, and identifying any anomalies
26 Naval ISR Fusion Principles, Operations, and Technologies

or deviations from normal patterns. Once an anomaly is detected, a situation

threat assessment is performed to assign a threat value to the situation which
enables decision makers to act as early and as distant from home as possible.
However, an efficient MDA strategy involves a continuous monitoring of ves-
sels within an area of interest or globally, as well as an improvement in the abil-
ity to collect, fuse, analyze, and disseminate information to decision makers as
shown in Figure 2.1.
The process of achieving MDA involves persistent monitoring of entities
in the global maritime domain ranging from keeping custody of the different
types of vessels (fishing vessels, leisure, cargo, military, etc.), to monitoring ves-
sel crews and passengers, to learning trends in the maritime routes and keeping
track of friendly and foe forces, to gaining situational awareness about ports
and other maritime infrastructures. This process involves the following steps
in order:

1. Data collection: MDA requires the aggregation of data from different

sources on people, vessels, facilities, infrastructure, and so on.
2. Data fusion: the availability to collect geospatial datasets makes moni-
toring the maritime domain feasible especially when integrated with
other forms of intelligence (signal intelligence (SIGINT), geospatial
intelligence (GEOINT), automatic identification system (AIS), and
open-source intelligence (OSINT)).

Figure 2.1 Conceptual model of the elements of the intelligence process [14].
 Principles of Operations from Seabed to Space 27

3. Data analysis: Data analysis conducted through machine learning al-

gorithms and artificial intelligence promises a faster and more autono-
mous tool to identify anomalous behaviors, trends, and patterns and
to predict behaviors.
4. Data dissemination: Once the data collected from different sources is
analyzed, decision makers are presented with an effective understand-
ing of the maritime domain either on a regional or global level (de-
pending on the mission) thereby enabling strategic, operational, or
tactical response.

Intelligence Source Terminology

Throughout this book we use standard intelligence terminology to

distinguish the many categories of sources, collection methods, and
analytic processes. The major recognized terms and acronyms are provided
as follows:

SIGINT is the exploitation and analysis of data derived from electronic

signals and systems used by foreign targets, such as communications
systems, radars, and weapons systems.
ELINT (electronic intelligence information) is derived primarily from
electronic signals that transfer data, but do not contain speech or text.
TechELINT is ELINT that focuses on describing the signal structure,
emission characteristics, modes of operation, and weapon system
associations of emitters.
OpELINT is ELINT that focuses on locating specific ELINT targets to
report the electronic order of battle (EOB) and threat assessments (or,
TacELINT).
COMINT is the collection and processing of foreign communications
(excluding open sources of communications such as radio and television
broadcasts) for the purpose of extracting information of value to
intelligence.
Cyberspace operations (CO) is the employment of cyber capabilities where
the primary purpose is to achieve objectives in or through cyberspace.
Cyber intelligence operations include:
Cyber operational preparation of the environment (C-OPE) is the
collection and analysis of information that maps the cyber environment
(e.g., networks, nodes, or channels);
Cyber ISR is the collection and analysis of cyber sources to support
offensive and defensive cyberspace operations.
28 Naval ISR Fusion Principles, Operations, and Technologies

GEOINT is the exploitation and analysis of imagery and geospatial

information to describe, assess, and visually depict physical features and
geographically referenced activities on the Earth. GEOINT consists of
imagery, imagery intelligence, and geospatial information.
HUMINT (human intelligence) is the collection and analysis of
information obtained from human sources.
ACINT (acoustic intelligence) is the collection and analysis of acoustics
signals to locate, identify, and track maritime vessels (generally from active
or passive SONAR systems) or overhead aircraft.
MASINT (measurement and signature intelligence) is the collection and
analysis of data obtained by quantitative and qualitative analysis of data
(metric, angle, spatial, wavelength, time dependence, modulation, plasma,
and hydromagnetic) derived from specific technical sensors for the purpose
of identifying any distinctive features associated with the emitter or source.
OSINT is the analysis of publicly available information (PAI) from many
sources, including foreign news sources, social media, and mobile device
location data.

ISR is a military operation that plays a pivotal role across all domains (air,
land, seas, space, and cyberspace) and enables the execution of MDA functions.
Its goal is not only to help decision makers and stakeholders “anticipate change,
mitigate risk, and shape outcomes,” [15] but it also has a great influence on how
missions and operational exercises are planned to defeat the adversary across all
domains. The U.S. DoD defines ISR as “an integrated operations and intel-
ligence activity that synchronizes and integrates the planning and operation
of sensors, assets, and processing, exploitation, and dissemination systems in
direct support of current and future operations” [16]. ISR platforms, that in-
cludes satellites, ships, humans, UAVs, UUVs, and planes, collect specific in-
formation using a wide range of sensors about an area of interest. The collected
data is analyzed with the help of AI and ML algorithms to not only produce
intelligence reports that support operational needs to defeat the adversaries but
also to protect the nation and its forces. Figure 2.2 illustrates the breadth of
ISR across all domains from seabed to space, including the cyber domain, to
detect objects of interest and predict their behavior within a battlespace. This
effort necessitates the fusion of all types of information (SIGINT, GEOINT,
MASINT, PAI, and HUMINT) across all domains and requires the collabora-
tion of all forces operating across the domains as well as working with allies to
ensure timely information flow from one service agency to another for a suc-
cessful execution of missions.
This concept is referred to as the Joint All-Domain Operations (JADO)
which is defined as “operations conducted across multiple domains and contest-
ed spaces to overcome an adversary’s (or enemy’s) strengths by presenting them
 Principles of Operations from Seabed to Space 29

Figure 2.2 ISR capabilities across all domains [17].

with several operational and/or tactical dilemmas through the combined ap-
plication of calibrated force posture; employment of multi-domain formations;
and convergence of capabilities across domains, environments, and functions in
time and spaces to achieve operational and tactical objectives” [18]. One of the
challenges that faces this concept is data handling. Standardizing the data across
services or even within each service can be challenging because each type of data
is stored in a specific format. Data is also stored at different classification levels
which makes data sharing a challenging task. The lack of a standard taxono-
my renders information exchange between different military and intelligence
services difficult unless a translation software is used to present the data to a
specific agency in its own lexicon. The availability of large amounts of data gen-
erated across all ISR domains necessitates the use of AI/ML algorithms where
humans and machines work together to process all this information. However,
one challenge in using AI/ML technology is the ability for decision makers to
trust the performance of the developed algorithms, which can only be gained
after long periods of test and development. This trust is further enhanced with
the algorithm’s demonstrated ability to detect drift, deception, and black swans.
On September 30, 2020, the deputy secretary of defense, Mr. David L.
Norquist, presented a DoD data strategy document that focuses on the need
to work with stakeholders in the operational domain and to “…treat data as a
weapon system and manage, secure, and use data for operational effect” [19].
Data needs to be:

• Visible: Users should be able to easily locate the data;

30 Naval ISR Fusion Principles, Operations, and Technologies

• Accessible: Users should understand the content of the data and the con-
text in which it was taken;
• Linked: Users should understand the relationships between different
data elements;
• Trustworthy: Users should trust the data in making decisions;
• Interoperable: Users across all domains should possess a common under-
standing of the data;
• Secure: Users should have confidence that the data is only available to
authorized users.

To ensure that the data satisfies the abovementioned goals, the following
capabilities need to be developed:

1. Architecture: Build an architecture that allows access to the data but

requires that capabilities be able to adjust in response to technological
and data requirement changes;
2. Standards: Employ standards for the managing, representing, sharing,
and utilizing datasets;
3. Governance: Data governance provides guidelines, processes, tools,
metrics, and policies needed to manage the data sets from the time
they are acquired to the time they are disposed;
4. Talent and culture: A shift towards a culture that revolves around the
data requires providing the necessary skills and training for the work-
force to take advantage of this new system to make data-based deci-
sions, create policies, and develop processes.

Some of the challenges facing MDA are:

1. Uncooperative vessels: Some vessels prefer to stay anonymous to avoid

detection and surveillance for several reasons (military, drug distribu-
tion, piracy, etc.). It is therefore a challenge to collect and analyze
information, which is often very sparse, on such vessels.
2. MDA enterprise establishment: Implementing a maritime enterprise ca-
pability that functions in a cohesive way to share information, analy-
ses, and maintains a situational awareness at all stages of a mission is
a major challenge. To determine anomalous behavior and understand
all maritime activities, different branches of the military need to coop-
erate within their services as well as with allies, commercial companies,
 Principles of Operations from Seabed to Space 31

and civilians. A successful MDA involves establishing an enterprise at

the international level. This necessitates the development of [20, 21]:
• MDA collaborative platform;
• Network to keep the MDA community connected: It is important
to keep MDA data accessible, agile, and flexible to all users to
address the real-time needs of the fleet to meet mission require-
ments and support decision makers;
• Quality control and quality assurance of MDA information: Pro-
tecting information collected from adversarial attacks is crucial to
ensure integrity and confidentiality of information.
3. Personnel information: For a successful MDA execution, it is impor-
tant to maintain security information on all personnel involved in the
maritime domain. This becomes challenging when considering all in-
volved parties in the maritime domain ranging from operators of fish-
ing vessels, cargo, and cruise lines to passengers.
4. Data acquisition: Situational awareness during peace and war times
requires detailed information to be gathered. This can be achieved
through acoustic, optical, and electromagnetic source methods using
different platform-based sensors (unmanned vehicles, satellites, so-
nobuoys, etc.). The data needs to address the need for real-time
analyses and dynamically changing mission plans to meet the
warfighter needs.
5. Resource orchestration: Allocation of resources to meet increased need
for surveillance on a specific area of interest calls for the design of an
agile and flexible infrastructure that allows for reallocation of resources
in a short time notice.

2.2 Naval Subsurface and Seabed Domains

In the subsurface domain, platforms include both manned and unmanned sys-
tems that leverage stealth in their execution of the ISR mission. Modern navies
deploy either manned nuclear or diesel-electric submarines, or a mix, classified
as attack, hunter-killer, or multipurpose/mission submarines. Typically, their
primary missions are special operations, ISR, and with the incorporation of
guided missiles, strike. Their ability to covertly collect intelligence and main-
tain station for extended periods of time make them premier ISR platforms.
Table 2.1 summarizes typical attack submarine ISR capabilities, the associated
phenomenology, and mission of each.
32 Naval ISR Fusion Principles, Operations, and Technologies

Table 2.1
Typical Attack Submarine ISR Capabilities
ISR System Phenomenology Mission Examples
SONAR system Propagation of Detection, acquisition, Thales UMS-3000 (UK),
(active, passive) sound energy identification, and localization AN/BQQ-10 A-RCI (U.S.),
of acoustic emissions and SONARtech Atlas
AI&R (Germany)
Electronic warfare Propagation of Detection, acquisition, AN/BLQ-10 (V) (U.S.)
system, electronic radio frequency identification, and localization
support measures (RF) energy of RF emissions
(ESM)

Unmanned underwater vehicles (UUV) also leverage stealth in the execu-

tion of the ISR mission, but with the added benefit of autonomy with no risk
to human life. Further, they extend reach into denied areas, and enable missions
in water too shallow or otherwise inaccessible for conventional platforms [22].
Current and proposed ISR missions include persistent and tactical intelligence
collection (signal, electronic, measurement, and imaging intelligence, meteo-
rology, and oceanography), chemical, biological, nuclear, radiological, and ex-
plosive (CBNRE) detection and localization, near-land and harbor monitoring,
deployment of leave-behind surveillance sensors or sensor arrays, specialized
mapping, and object detection and localization [23]. UUV systems can be cat-
egorized into four classes according to their size. Sizes range from man-porta-
ble vehicles with payload capacities of less than 0.25 ft3 to large vehicles with
payload capacities of 15 to 30 ft3 with the possibility of external stores. Light
(LWV) and heavy weight vehicles (HWV) find utility as well. Based on size,
weight, and power constraints, each class of UUV offers a unique contribution
to the ISR mission as summarized in Table 2.2 [23].

Table 2.2
Classes of UUV, Their Nominal Levels of Performance, and ISR Applicability
Endurance Endurance
Diameter Displacement High Hotel Hotel Load Payload ISR
Class (inches) (lbs) Load (hours) (hours) (ft3) Application
Man- 3–9 <100 <10 10–20 <0.25 Special
portable purpose
LWV 12.75 ~500 10–20 20–40 1–3 Harbor
HWV 21 <3,000 20–50 40–80 4–6 Tactical
Large >36 ~20,000 100–300 >>400 15–30 with Persistent
external
stores
From: [23].
 Principles of Operations from Seabed to Space 33

Undersea surveillance sensors also contribute to subsurface domain ISR

capabilities. These systems are deployed as fixed or mobile platforms. Gener-
ally, fixed platforms are passive SONAR systems that provide a persistent ISR
capability by the employment of large, fixed arrays of SONAR hydrophones to
detect and identify maritime traffic, particularly submarines, at strategic ingress
and egress locations. The mobile platforms are passive SONAR systems that are
towed from a surface platform. Similarly, the mission is to detect and identify
vessels of interest, particularly submarines. A summary of typical naval undersea
surveillance capabilities is provided in Table 2.3.

Table 2.3
Active and Passive SONAR Systems that Contribute to ISR
System Description Example
Surface ship A passive acoustic surveillance system towed U.S. Navy UQQ-2
surveillance towed from T-AGOS surface platforms. It provides passive
array sensor system detection of nuclear and diesel-electric powered
(SURTASS) submarines, and real-time reporting to theater
commanders and operational units [24].
Surface ship Active adjunct to the SURTASS SONAR system U.S. Navy WQT-2
SURTASS / low [26]. Also compact low-frequency active (CLFA) for
frequency active (LFA) SURTASS.
Surface and SONAR arrays (active and passive) mounted on ship Various
submarine hull- and subhull SONAR domes.
mounted arrays
Surface air-deployed Patrol aircraft drop short-lived (hours) SONAR AN/SSQ-36B, AN/
expendable buoys (active and passive) in grid patterns that SSQ-53G, AN/SSQ-
sonobuoys locate submarine (and surface ship) signatures and 62F [25]
transmit data back to the aircraft. The sonobuoy
deploys a cable-suspended hydrophone sensor.
Advanced multisensor units may employ e-field and
magnetic sensors.
Surface advanced Massive arrays of station-keeping and floating In development by
sensor grids sensors are networked and transmit processed Defense Advanced
reports when needed via satellite. Sensors collect Research Projects
environmental data (ocean temperature, sea state, Agency (DARPA)
and location) as well as acoustic activity data about Ocean of Things
vessels and aircraft. (OoT) [26]
Seabed fixed A network of hydrophone arrays mounted on the U.S. sound
surveillance system seafloor that provides to ashore centers threat surveillance system
(FSS) location information to provide an accurate (SOSUS), fixed
operational maritime picture of submarine and distributed system
surface traffic. (FDS), and FDS-
commercial (FDS-C)
Seabed advanced Rapidly deployable undersea surveillance system Lockheed ADS [27]
deployable system that uses distributed passive acoustic bottom
(ADS) mounted arrays wirelessly linked to an analysis and
reporting system to provide continuous acoustic
coverage over large areas of the ocean.
From: [28, 29].
34 Naval ISR Fusion Principles, Operations, and Technologies

2.3 Naval Surface and Airborne ISR

Naval surface ISR capabilities include both manned and unmanned platforms.
Manned surface ISR systems have become highly integrated with combat sys-
tems due in large part by the increasing adoption of network-centric operations.
Thus, sensor suites for target acquisition find utility in informing situational
awareness, cueing other sensors, and supporting other support to ISR and vice
versa [30].
Unmanned surface systems, known as unmanned surface vessels (USV),
have several distinct advantages over their manned counterparts. They enable
access to areas where the risk of using manned platforms is unacceptably high,
they provide extended range and persistence on station, and are expendable at
mission termination [31].
Airborne ISR platforms offer range and speed advantages over other con-
ventional ISR platforms like those in the subsurface and surface domains. Naval
airborne ISR capabilities include both manned and unmanned platforms. Table
2.4 is a summary of typical naval airborne ISR capabilities with a brief descrip-
tion and examples of each.

Table 2.4
Typical Naval Airborne ISR Capabilities
Aircraft Class Description Example
Multi-intelligence Multi-intelligence mission employing radar, EP-3E Aries II (U.S.)
reconnaissance sonobuoys signals intelligence and other sensors
integrated with offboard systems

Maritime patrol Maritime missions, roles include search and rescue, P-8 Poseidon (U.S.),
antisurface, and antisubmarine missions Tupolev Tu-142 (RUS)
Fighter Multirole attack and fighter aircraft F-35 Lightning (U.S.),
Sukhoi Su-27 (RUS)
Electronic warfare Electronic attack mission with onboard signals EA-18G Growler (U.S.)
collection
Early warning and Airborne command and control, battle space E2-HawkEye (U.S.),
control management Beriev A-50 (RUS),
KL-2000 (PRK)
Rotary wing ASW, anti-surface warfare (SUW), and SH-60 Seahawk (U.S.)
electromagnetic warfare (EW), combat search and
rescue (CSAR), organic airborne mine countermeasure
(AMCM), and combat support missions
UAV Persistent maritime ISR MQ-4C Triton (Triton)
Small UAS Over-the-hill short range ISR RQ-11B Raven (U.S.)
High altitude and Manned and unmanned high-altitude platforms with U-2, Global Hawk [35]
endurance imaging radar, EO, and signals intelligence sensors
From: [33].
 Principles of Operations from Seabed to Space 35

2.4 Naval Space

Why are navies across the world interested in space? Space allows the military
to transit objects such as ballistic missiles and to launch satellites at different
altitudes for the purpose of ISR or communication applications. Space-based
sensors collect data that is crucial to execute tactical missions and military op-
erations. Space offers a unique domain for the military to exploit because it is
characterized by [34]:

• Global perspective: Space assets, when working effectively together, allow

the warfighter to gain a worldwide coverage and a global overview that
enables navies and military services to obtain a space, maritime, and
land situational awareness, assess risk, and prevent adversarial attacks.
• Rapid revisit times: Low-Earth orbit (LEO) spacecrafts orbit the Earth
every 90 to 120 minutes and are characterized by fields of view of about
hundreds of miles. This enables the tracking of objects or surveilling
areas of interest.
• Persistence: Geosynchronous satellites complete one orbit around the
Earth every 24 hours so that the satellite appears stationary with respect
to the Earth and therefore can observe the same location on Earth all
the time. These satellites also provide weather forecasts that deem to be
useful in planning military operations and making predictive analysis
involved in ISR missions.
• Access to denied areas: Unlike the air domain which is governed by a set
of rules that can restrict access to specific air spaces, there are no laws
specifying territorial boundaries in space thereby making it possible for
navies and other military organizations to use space to enable ISR and
remote sensing missions.

Figure 2.3 shows the different altitudes at which satellites orbit, along
with domains that are most useful for naval ISR.
The use of space for military operations began towards the end of World
War II when Germany used ballistic missiles to strike England [39]. With the
advancement of space capabilities and the decrease of their relative cost, space
capabilities now provide a major support to military forces around the globe.
This includes providing warnings against missile attacks, the ability to geolocate
and identify objects of interest, precise navigation, and tracking of adversary
activities. When satellites are launched into space, they are assigned a specific
task to achieve. Therefore, one can classify satellites based on their functions.
36 Naval ISR Fusion Principles, Operations, and Technologies

Figure 2.3 Types of satellite orbits and their applications [35].

Table 2.5 provides a summary of the major categories and functions of satellites
along with a few examples [18, 19, 37].
The advancement of space-based capabilities is making it very challeng-
ing for nations and groups to hide their military operations from competing
nations. As a result, there are multiple threats associated with space [18, 38]:

• Cyberspace threat: Cyberspace, which will be discussed in detail in Sec-

tion 2.5, poses a major threat to the operation of space assets because
with a comprehensive knowledge of the data network and the satellite
command and control system, adversaries can attack the space system
infrastructure.
• Directed energy weapon: This technology uses focused energy sources
such as lasers or high-power microwaves to discreetly damage a target,
since radiation outside of the visible spectrum is invisible and does not
produce sound.
 Principles of Operations from Seabed to Space 37

Table 2.5
Major Satellite Functional Categories to Support Naval Missions
Satellite Functions Function Explanation
Communication Communication satellites allow for:
satellites ·· Voice communication;
·· Broadband internet;
·· Data transfer.
Positioning, Constellations of navigation satellites allow for the determination of
navigation, and timing location as well as time using global or regional satellite navigation
satellites systems.
Remote sensing Remote sensing satellites monitor the Earth to track resources that are
satellites essential to human wellbeing. This includes weather monitoring and
prediction, natural disaster damage assessment, animal migration, and
agricultural monitoring.
ISR satellites ISR satellites provide signal intelligence, imagery intelligence, and an
overall global situational awareness about an area of interest. There are
four types of ISR satellites:
·· Electro-optical imaging;
·· Radar imaging;
·· Signal intelligence;
·· Communication relay satellites.
Weapon delivery Satellites are used to track missiles and hypersonic weapons to ensure a
satellites fast and precise weapon delivery.
Search and rescue These satellites have the capability to detect and locate vessels, aircraft,
satellites and individuals when in dangerous situations or in remote places by
receiving beacon transmissions from a small transmitter.

• Electronic warfare: This technique involves the use of electromagnetic

energy to disrupt the RF transmission to or from satellites and control
the spectrum to deny the adversary access the electromagnetic spectrum.
• Kinetic energy threats: Kinetic energy threats, also known as antisatellite
(ASAT) missiles, are designed to physically destroy satellites from a dis-
tance. This is done using a projectile, a missile, launched from a station-
ary point on the Earth, sea, or from an aircraft. There are other methods
to destroy satellites such as releasing a cloud of pellets in the trajectory of
the satellite or detonating an explosive device near a satellite. However,
using ASAT is the technique that is most developed and tested due to its
precision in hitting the target.
• Orbital threats: This technique consists of launching a satellite into an
orbit for an extended time before it engages in attacking the target satel-
lite. Such a system could use a jamming technique, directed energy, or
chemical sprayers to interfere with the satellites.

Space sensors provide a broad view of the oceans, ports, and canals that
vessels traverse and provide the potential to contribute persistent monitoring
38 Naval ISR Fusion Principles, Operations, and Technologies

of individual vessels and fleets in blue water as well as the littorals. Space con-
stellations are groups of multiple satellites (similar or diverse) that operate as a
system to perform a designated mission, such as maritime surveillance. Naval
ISR requirements that are fulfilled by space constellations include:

• Periodic monitoring and tracking of commercial and naval vessels coop-

eratively using the internationally mandated AIS reporting transponders
on ships;
• Monitoring ports, canals, and waterways for vessels at near-shore opera-
tions;
• Location of vessels that have gone dark by disabling legally required AIS
signals and by controlling all emissions (EMCON);
• Detecting, locating, and positively identifying noncooperative vessels of
interest.

These requirements apply to naval, intelligence, and law enforcement

missions to monitor the world’s maritime traffic.
Consider three major commercial categories of sensors that can contrib-
ute to maritime vessel tracking (Table 2.6):

• Synthetic aperture radar (SAR) constellations provide all-weather ship de-

tection with a wide area of coverage; constellations of multiple satellites
provide revisit rates that have the ability to maintain custody of slow-
moving (and nonevading) maritime traffic. SAR provides high-resolution
images both day and night, independent of weather conditions [39].
• EO imaging constellations provide port and ship imaging with sufficient
ground sample distance (ground sample distance (GSD), a measure of
image resolution) to distinguish ship classes (e.g., merchant, combatant,
guided missile destroyer, or small craft). Lower resolution sensors (GSD
0.5m to 1m) perform ship class discrimination and higher-resolution
sensors (GSD <1m) may even discriminate an individual ship hull and
name [40]. Unlike SAR, the EO sensor cannot operate at nighttime (un-
less an infrared (IR) band is added) and is blocked by heavy cloud cover.
For this reason, SAR is the first choice for locating vessels and then EO
may cross-cued to observe and verify the identification of a ship when
it is feasible [41].
• Radio frequency (RF) satellite constellations monitor ship-reporting sen-
sors and other emissions from the AIS and other vessel emissions. AIS
is the global standard for ship identification between ships and shore
trackers, and for global at-sea monitoring by satellite collection of the
 Principles of Operations from Seabed to Space 39

Table 2.6
Representative Commercial Spaceborne Maritime Surveillance Systems
Capability Example System Constellation Parameters Sensor Parameters
SAR Radarsat ·· 3 SAR satellite vehicles ·· C-band;
Constellation ~600-km LEO; ·· High-resolution 1m × 3m
Mission [44] ·· 4-day exact revisit time; (spotlight mode);
·· 30-minute data latency from ·· Medium resolution ship
acquisition to delivery; detection mode resolution
·· 15 minutes of imaging time 50m × 50m; swath width
per orbit per satellite. 350 km.
IceEye [45] SAR ·· 18 SAR satellite vehicles ·· X-band;
(planned 2022) ~600 km LEO; ·· Resolution in modes strip =
·· 1- to 22-day revisit time 3m, spot = 1m, and scan =
depending on satellite and 15m.
target location.
Capella SAR ·· 36 SAR satellite vehicles ·· X-band;
(planned 2023) ~600 km LEO; ·· Highest resolution 50 cm.
·· 2- to 4-hour revisit time
depending on satellite and
target location.
Electro- MAXAR ·· 2 satellites; ·· High-resolution 30–50 cm
optical (EO) Worldview ·· 617-km LEO sun synchronous; optical imagery.
·· 2-day revisit.
Planet Skybox [46] ·· 15 satellites; ·· Color and near-infrared (NIR)
·· LEO. imagery;
·· 90-cm resolution.
Planet Dove [47] ·· 88 satellites; ·· Bayer-masked charged-
·· International Space Station couple device (CCD) camera;
orbit, sun synchronous. ·· Ground sampling distance
(Nadir) ranges between 2.7m
to 4.9m;
·· Spectral band: red, green,
blue, NIR;
·· 3m multispectral image
resolution.
RF signal ORBCOM (AIS) ·· >30 satellites; AIS receivers are aggregated
cooperative ·· LEO; sun synchronous evenly to provide continuous vessel
vessel tracking spaced vehicles. tracking via subscription
service
Spire (AIS; ADS-B) ·· >100 nanosatellite vehicles; SDRs for maritime AIS and
cooperative ·· LEO; 500-km sun aviation ADS-B tracking.
vessel tracking synchronous.
HawkEye 360 (RF) ·· >6 clusters of 3 formation- SDRs geolocate VHF marine
noncooperative flying satellites; radios, UHF push-to-talk radios,
geolocation of ·· LEO, 575 km. maritime radar systems, AIS
emitters beacons, and L-band satellite
devices.
Note: SDR is software defined radio; ADS-B is automatic dependent surveillance-broadcast.

periodic AIS signals [42]. AIS automatically provides vessel informa-

tion to other ships and coast authorities fitted with an AIS transponder
40 Naval ISR Fusion Principles, Operations, and Technologies

that operates in the very high frequency (VHF) mobile maritime band.
The data provided by AIS is spatiotemporal data that consists of lati-
tude, longitude, course-over-ground, speed-over-ground, vessel identi-
fier number, navigational status, time stamp, and so on. AIS reporting
contributes to vessel tracking for MDA as well as for collision avoid-
ance, and search and rescue. AIS is a cooperative system that requires the
cooperation of vessels and therefore does not account for dark vessels.
Those are vessels that evade tracking by turning off the AIS transponder
or by providing incorrect ID codes or locations. Of course, maritime ra-
dars and other emitters on ships provide potential sources for detection
and location by RF constellations that use interferometric methods to
localize the emitter location [43].

In each case constellations of multiple satellites with sensors contribute

maritime information; the correlation and combination of this data from all
three modes (data fusion) can enhance the ability to monitor and track the
movement of individual vessels. While we provide examples of commercial
satellite constellations, dedicated military or naval constellations can directly
provide this information for intelligence, surveillance, reconnaissance, and tar-
geting (ISRT).

2.5 Naval Cyber

Cyberspace, often referred to as the fifth operational domain, refers to a do-
main that hosts interconnected networks of data and the information contained
within those networks. As navies operate in a global environment, they are
exploiting large amounts of data collected by systems deployed from seabed to
space, they are fusing all this information, they are sharing it, and they are com-
municating it through sophisticated networks to achieve situational awareness
and military dominance. As a result, the military does not only depend on the
cyberspace domain to execute specific cyberspace operations, but it relies on
cyberspace for the success of military missions and operations in the sea, land,
air, and space domains. The cyberspace is characterized by the following [48]:

• Innovation: Cyberspace evolves in response to advances in technology,

system processes, and architectures to produce capabilities that meet
the needs of the military. Therefore, it requires constant vigilance and
awareness of the most recent technological breakthroughs that affect the
domain.
 Principles of Operations from Seabed to Space 41

• Complexity: One way to understand part of the complexity of the cy-

berspace is by focusing on the geophysical hierarchy of all domains. The
surface of Earth (land domain) is surrounded by the maritime domain,
which is surrounded by the air domain, which in turn is surrounded by
the space domain. However, the cyber domain is embedded in all four
domains and mission success in all these domains depends on a success-
ful cyberspace operation.
• Interdependence: Cyberspace is formed by an interconnection between
information, and the hardware and software systems that store, analyze,
and transmit the data.
• Volatility: Because the cyberspace keeps changing, it can be a challenge
to ensure effectiveness of some cyber operations and ensure that vulner-
abilities are not introduced because of such changes.
• Speed: Information within the cyber domain moves quite rapidly to al-
low commanders and warfighters to make decisions and conduct opera-
tions. The cyber domain becomes increasingly effective as the speed for
processing all the collected data and analyzing it to make operational
and strategic decisions improves.

Cyber superiority is achieved by realizing an infrastructure that:

• Provides freedom to the stakeholders to access, analyze, manipulate, and

communicate information between services and across all domains.
• Denies access of the cyberspace to adversaries during specific times and
at certain locations.
• Protects the network by ensuring the security of each of the different
layers that form the cyberspace domain [49]:
1. The physical layer: This layer consists of physical and geographical
components, the hardware backbone of cyberspace. A prominent
and obvious example of a physical layer is the fiber-optic network
of cables including undersea cables. Another example includes the
constellation of satellites.
2. The logic layer: This layer involves the data in the state of rest as well
as in a state of transmission. It is the cyberspace analog to the central
nervous system. The logic layer is responsible for retrieving files and
for sending and receiving messages. Much like the human brain, the
logic layer is where decisions are made. Two key elements of the logic
layer are domain name servers (DNS) and Internet Protocols (IP).
42 Naval ISR Fusion Principles, Operations, and Technologies

3. The information layer: This is the layer that acts as persona part of cy-
berspace and includes examples from the internet, such as websites,
chats, emails, photos, documents, and apps. It represents individuals
and organizations, both real and fictitious. Without the previous two
layers, the information layer cannot function.
• Offers the ability to perform offensive operations, the ability to gain
information with high intelligence value by:
1. Attacking the cyberspace by performing actions or manipulations
that create noticeable denial effects;
2. Exploiting the cyber domain of adversaries to gain intelligence.

Naval fleets operating afloat in the maritime environment coordinate cy-

ber activities with ashore cyber units to conduct the operations in three areas
across the global network (Table 2.7). Because the afloat cyber operators con-
duct activities that are closer to adversary fleets they may have enhanced capa-
bilities to access and monitor target networks.
Cyber ISR operations include a wide range of monitoring activities on
internet networks, dark networks, and social media channels to perform threat
assessments for afloat commanders.

2.6 A Seabed-to-Space Scenario

In this section we provide a scenario that describes the roles of the various naval
platforms and sensors in a representative situation to illustrate how networked
sensors provide ISRT to enable naval operations. The scenario is designed to
demonstrate, in a near-realistic situation, the many variables that come into
play to conduct naval operations that demand excellent ISRT and maritime
awareness.

• The scenario is a sequence of imagined events that unfold in a naval

operation;
• A naval operation is a naval action (or the performance of a naval mis-
sion) that may be strategic, operational, tactical, logistic, or training. It
is a process of carrying on or training for naval combat to gain the objec-
tives of any battle or campaign [51].

The brief scenario is represented by the following elements:

• Background and situation are the starting geopolitical and military (na-
val) events that drive the need for naval operations;
 Principles of Operations from Seabed to Space 43

Table 2.7
The Categories of Naval Cyber Operations
Preparation and ISR Defensive Offensive
·· C-OPE; ·· Defensive counter cyber ·· Offensive cyberspace
·· Network ISR. (DCC); operations (OCO);
·· Proactive NetOps, defensive ·· Offensive counter cyber
countermeasures. (OCC).
·· Map target networks; ·· Follow established defense ·· Apply offensive tools, tactics,
·· Assess adversary cyber order policies, techniques, tactics, and procedures;
of battle; and procedures (TTPs); ·· Conduct offensive operations
·· Monitor threatening conduct operations to: as required under proper
activities; ·· Monitor and guard authorities;
·· Provide indications and processes to prevent and ·· Synchronize ashore-afloat
warning; delay attackers. cyber ops with other noncyber
·· Develop targeting ·· Report and restore cyber operations;
information, perform strategic incidents. ·· Perform effects assessment;
target development. ·· Establish persistent presence
as required on threat
networks.
From: [50].

• Geospatial area description describes the land-ocean configuration and

constraints of the topography from ocean floor to littoral sea-land coast-
al areas and the terrain of land features;
• Order of battle of the Red and Blue opposing forces, and the numbers of
platforms and their locations and objectives; this drives the threats (to
both forces) and naval objectives;
• Scenario progression is the series of actions that both naval forces take over
time and the events as they interact. This interaction includes conflicts
in the physical-kinetic, electronic, cyber, and political-social domains.

Background and Situation

This scenario takes place in the fictional Kandago Sea where two nations have
contested claims over a series of 10 islands across a 400-km chain in the Kan-
digan Straits. The dispersed islands are contested based on complex historical
claims and counterclaims due to the nations’ authority in their territorial seas
(out to 12 nautical miles) and jurisdiction in the exclusive economic zones (out
to 200 miles) that intersect at the minor islands. The islands had little value un-
til recent decades, when their proximity to potential oil and mineral resources,
and their strategic location in the path of major shipping lanes have increased
the level of geopolitical competition. The islands provide strategic positions to
monitor the sea lanes and enable limited naval presence to control or harass
shipping. In addition, they can provide fueling, logistics, and command and
control to enable future conflicts with the Green Island. The largest island has
44 Naval ISR Fusion Principles, Operations, and Technologies

terrain suitable for a military air strip enabling a strategic placement for air
surveillance and strike [52]. Green island nation has been the historical owner
of the islands and has been concerned about losing sovereignty over the chain
within its declared sovereign territorial waters.
This scenario takes place over a 4-day period. Red nation political rhetoric
and air reconnaissance over the first 2 days (days –2, –1) provide indications
and warnings (I&W) of potential Red nation intent to annex one or more of
the islands in the chain. This warning allows Blue forward presence naval forces
to move toward the area and to use air capabilities to deploy unattended subsur-
face and seabed sensors. The Blue Surface Action Group (SAG) is conducting
maritime security operations (MSO) and is immediately called to move toward
the crisis [53]. The Blue force may be called upon to conduct littoral operations
in a contested environment (LOCE) if nonkinetic operations cannot mitigate
the aggression. Commanders’ intents, based on strategic orders are:

• Red force commander’s intent: Swiftly position sufficient land forces and
weapons to secure two strategic islands (Kandigan A and Kandigan B)
before military, diplomatic, and regional pressures can dislodge the force
and require a retreat to port. Defend the landing forces with SUW by
denying defending forces access to the straits, and only resort to land-
based strategic antiship missiles as a last recourse.
• Blue force commander’s intent: Respond quickly to support treaty ally
and economic partner Green nation in repelling Red nation aggression.
Move to position and stall annexation landing operations by presenting
a military threat sufficient to allow regional political powers to force
Red nation to cease landings and return to base. Threaten long-range
counter-naval operations and the ability to perform expeditionary ad-
vanced base operations (EABO) if necessary, to blunt any attempted
annexation of the island chain [54]. (This intent requires immediate
ISRT to identify and track the Red SAG to prepare for surface action
and expeditionary landings if necessary.)

Geospatial Area Description

The conflict arises in an approximately 200 × 150-km area of contested straits
between the Red mainland nation and the Green island nation (Figure 2.4).
Red naval forces emerge from their naval port to insert landing ships on two
of the contested islands (A and B) that are lightly populated by Green fishing
colonies. Blue forces moved from Blue water operations to enter the Kandigan
Straits from the east. Red forces can move from their naval base, loiter and as-
semble into an action group, and then traverse the 150 km to the target islands
within 4–6 hours. The Blue force, conducting MSO 100 km to the east will re-
Principles of Operations from Seabed to Space 45

Figure 2.4 Geospatial area and forces laydown.

46 Naval ISR Fusion Principles, Operations, and Technologies

quire several hours to arrive at the straits, and in-range to conduct organic ISRT
and kinetic operations [55]. The Blue force of 10 vessels will remain distributed
across the straits and back into the blue water to avoid the Red force advantages:
numerical at-sea vessel superiority, seabed sensor networks to Red ashore C2
centers, likely ISR assets on the island chain, and land-based missile threats.

Order of Battle
Red and Blue have SAGs (temporary or standing organization of combat-
ant ships, other than carriers, tailored for a specific tactical mission) that are
equipped as compared in Table 2.8.

Scenario Progression
Over a 2-day period (days –2, –1) Red nation provides I&W of the Red nation
intent to annex one or more of the islands in the chain. Political rhetoric ratch-
eted up the accusations of Green nation’s over-fishing the straits and unfounded
territorial claims. This warning causes regional partnership nations to call on
Blue nation to prepare to move Blue forward presence naval forces toward the
straits.
The implication for ISRT activities is that the Blue SAG and supporting
Blue nation capabilities must rapidly move from peacetime intelligence prepa-
ration of the operating environment to surveillance-reconnaissance (Table 2.9).
This rapid shift requires the immediate refocus of space and other sources to the
Blue port, SAG platforms, and potential targets of the aggression. As the SAG
moves to the area of operations, it deploys organic ISR capabilities to focus on
the threatening SAG. Reconnaissance is performed by unmanned and manned

Table 2.8
Order of Battle
Red Naval Forces (Aggressors) Blue Naval Forces (Defenders)
Land Force: Land Force: None
3 batteries, medium range (400 km) tactical
antiship missiles.
·· 5 DDG guided missile destroyers (1 additional ·· 3 DDG guided missile destroyers in SAG, 1
in reserve in port); Helo each;
·· 4 landing ships with 3 Helo each, and small ·· 2 expeditionary sea base ships with EABO
vessels (2 in reserve in port); marine battalion, 4 Helo, and small vessels;
·· 2 FFG-guided missile frigate; ·· 2 FFG-guided missile frigate;
·· 3 aux support ships; ·· 2 auxiliary support ships;
·· 2 tactical attack submarines. ·· 1 tactical attack submarine.
Air Capability: Air Capability:
·· 12 long-range patrol aircraft with antiship ·· 5 UAV reconnaissance drones (long range);
missiles. ·· 12 UAV reconnaissance drones (short range).
 Principles of Operations from Seabed to Space 47

Table 2.9
Activities Across the Phases of ISRT Operations
ISRT JIPOE Situation Surveil
Phase Intelligence Awareness Recon Targeting Fire Control
Cyber Cyber monitoring and initiate target access Coordinated cyber node attack
Space Orchestrate for aggregate Focus Focus on Support
coverage per maritime collection orchestration
target coordinated fires
priorities on combatants
priorities
Air Normal fleet coordination of manned and Coordinated Coordinated air-
unmanned sensor assets; air-surface surface fires
Surface Adapt deployed sensors, relay comms, to target
predicted threats, threat axis. tracking
nomination
Subsurface Coordinate manned, unmanned, and relay comms ASW and SUW operations
for sensing, maneuvering
Seabed Monitor-report early precursor Focus seabed Issue and relay commands
adversary activities sensors; move to enable sensors and target
deep UUVs to mines, torpedoes
position
DAYS → HOURS → MINUTES

air capabilities to deploy unattended subsurface and seabed sensors, reconnoiter

the SAG platforms, and extend the SAG’s signals collection capability.
The following paragraphs summarize the events of days 1 and 2 that bring
the scenario to resolution. This is a contrived result to illustrate exceptional
ISRT capabilities, but it is but one of many possible outcomes, many of which
do not result in Blue force success. The fictitious scenario is compressed in time
and orchestrated to illustrate the dynamic and interactive role of ISRT.
Day 1. After 48 hours of building political rhetoric and air reconnaissance
over islands A and B, Red initiates the aggression with combatant and support
ships departing port and forming the SAG to head toward the target islands
A and B. Guided missile destroyers (DDG) deploy to defensive positions to
guard the landing force and prepare for surface action if (and when) Blue SAG
responds. The actions of Red and Blue forces on day 1 are summarized in Table
2.10. Blue force initiates focused ISRT and plans to move quickly from intelli-
gence preparation to surveillance-reconnaissance and then targeting. The ISRT
actions performed by Blue include:

1. Exploit commercial and national space imagery to locate the Red SAG
force; identify ships moved from port and hull numbers. Prepare for
the challenge of tracking, losing, and re-acquiring the SAG as it moves
across dense commercial ship lanes to its targets. Task commercial
SAR satellites to collect on areas that traverse the shipping lanes.
48 Naval ISR Fusion Principles, Operations, and Technologies

Table 2.10
Day 1 Actions
Period Red Actions Blue ISRT Actions
0000–0600 Order to execute plan KANDIGAN Receive ISR report—Red assembling SAG;
45; Monitor ISR reporting, plan organic ISR upon
Move ships from port; assemble arrival.
SAG to move.
0600–1200 Begin movement to target islands; Monitor cyber (social media) and EW channels;
Initiate EMCON (emission control); Heighten cyber alert status to prosecute and
Disable AIS transponders. respond to potential Red cyber incidents or
events;
Launch organic UAV ISR to monitor Red SAG.
1200–1800 Arrive and loiter 10 km off-islands: Arrive in standoff position 40 km off islands;
·· Issue warnings via radio and Monitor cyber and EW Channels, jam selected
social media. comm. channels, and coordinate with ashore
SAG ready to make land at 2400. cyber operators.
1800–2400 Stand by to move to 3 km for Maneuver to demonstrate ready-for-surface
landing ships; engagement;
Insert special landing recon Launch UAV aircraft to monitor Red landing
personnel; activities with IR night vision full motion video
Await 1200 attack orders. sensors.

2. Analyze topographic maps, bathymetry of the Kandigan island chain

to estimate likely target island(s) and landing sites. Direct available
ISR collection to those areas to support any Blue marine expedition-
ary ground activities. Identify any potential for Red mining operations
and requirements for countermining.
3. Deploy UUVs from ships, SONAR sensors (surface) from UAVs, and
task submarines to predeployed seabed SONAR sensors to locate Red
submarine threats and distinguish surface combatants from the dense
commercial traffic. Track DDGs and plan counternaval operations to
prepare to engage the surface force. Initiate persistent unmanned air
vehicle (UAV) coverage of the most likely target islands and combat-
ants.
4. Monitor civilian channels in social media (both in Red and Green) to
aggregate popular support/opposition for Red naval activities.
5. Receive strategic guidance updates from Blue ashore national com-
mand authority as regional diplomatic interactions place pressure on
Red political leadership.

Based on the ISR results, Blue prepares to perform countertargeting (us-

ing deception and decoys) to reduce the ability of the Red SAG to target the
Blue SAG as it moves within range of the land-based antis-ship missiles and
Red DDG missiles. (Of course, Red wants to avoid a surface engagement with
 Principles of Operations from Seabed to Space 49

Blue due to political blowback; its objective is to quickly complete and secure
its landing mission before Blue can disrupt the operation.)
Day 2. Through the night of day 2 (Table 2.11), Blue forces maneuver to
increase their ISR footprint and defensive posture, Blue submarine and UUVs
are maneuvered to detect and hunt the Red submarine threat. Red forces initi-
ate landing operations for A and B at 0530 with the objective of securing the
two islands by 1500 to enable a Red press conference announcing annexation of
“Restored National Lands to Protect Fishing” by 1600. At 0730, Red initiates a
social media campaign propagating the meme that Green mainland fishermen
have stepped up the harassment of the fishing villages of “Red National Lands
on the Islands.” The message is picked up and propagated by Red regional
news sources for morning news reports. By 1000, the situation escalates as Red
landing forces are headed ashore, but are hampered by precision electronic war-
fare jamming of their communication network and cyber disruption of their
afloat networks. Blue has maneuvered DDGs to coordinate unmanned air, sur-
face, and subsurface vehicles while poising a potential threat to Red DDGs by
long-range cooperative engagement if necessary. Blue increases cyber and EW
activities to slow the C2 of the landing effort and feigning the landing of the
EABO Marine battalion by helicopter and small vessels on island B to cause
Red to consider implications of a regional conflict. By 1300, the landing is only

Table 2.11
Day 2 Actions
Period Red Actions Blue ISRT Actions
0000–0600 Initiate landing on schedule; Maneuver DDGs to coordinate unmanned
Track Blue force maneuvering. air, surface, and subsurface vehicles;
Prepare for long-range cooperative
engagement of Red DDGs if necessary;
Increase cyber and EW denial, disruption
activities.
0600–1200 Provide C2 for landing, supply, and Feign the landing of the EABO Marine
defensive support DDG vessels; battalion by helicopter and small vessels
Incur stiff jamming environment and on island B;
reduced comms. Receive approval to expand offensive
cyber actions against Red port facilities
and Red patrol aircraft.
1200–1800 C2 unable to fully coordinate and monitor Reduce active engagement of cyber and
forces; RF channels to monitor withdrawal.
C2 detect imminent Blue marine force
arrival;
Withdraw forces.
1800–2400 Continue disinformation campaign that Maintain unmanned vehicles on station
Red forces protected island fishermen to provide ISR of Red forces returning
from Green harassment, while Blue to port;
forces carelessly aggravated the Monitor social media information
situation by threatening Coast Guard campaigns and measure effectiveness.
operations.
50 Naval ISR Fusion Principles, Operations, and Technologies

partially completed and C2 is sufficiently confused that Red national command

cannot determine the effectiveness of the operation; Red political leadership
calls off the operation and issues a statement that “Red forces protected island
fishermen from Green harassment, while Blue forces carelessly aggravated the
situation by threatening Red Coast Guard operations.” Supporting social media
campaign carries the meme with untrue anecdotal accounts of injured fisher-
men, Blue force landings on islands, and a Blue force missile launch against the
Red fleet that failed in-flight.
The ISRT actions performed by Blue include:

1. Receive strategic guidance updates from Blue national command au-

thority as regional diplomatic interactions place increased pressure on
Red political leadership using ISR-supplied intelligence to inform Red
of regional knowledge of Red’s actions.
2. Locate subsurface threats; UUVs employed active SONAR to locate
the Red submarine threat. They also pinged seabed sensors to “report-
on-detection” via deployed once-shot float-to-surface relays that relay
information via satellite communication.
3. Red fleet tracking: Blue space constellations (SAR imaging) had been
tasked to monitor Red forces throughout the night and detected the
landing assembly and preparations at 0500; these detections cued
UAV ISR to track the landing activities. The constellations also track
Red DDGs and provide updates to allow organic UAVs to track ma-
neuvers and provide relay if cooperative surface-to-surface missile en-
gagements are required.
4. Monitor and perform traffic analysis on Red frequency-hopping com-
munication and C2 channels to support jamming operations; analyze
the traffic patterns to distinguish operational phases. Blur mapped
cyber network data to engage Red cyber networks that support the
landing activities.
5. Continue to monitor civilian channels in social media (both in Red
and Green) to measure the impact of Red information campaigns;
conduct a counterinformation campaign to targeted channels to miti-
gate the effects of misinformation.

Scenario Conclusion
This contrived scenario provides a basic insight into the large number of vari-
ables, the range of alternatives, and the complexity of choosing naval tactics and
applying ISRT to achieve operational objectives. The role of ISRT is shown to
be a critical component to enable maritime awareness, insightful choice of op-
 Principles of Operations from Seabed to Space 51

erational actions, and precision in conducting offensive operations from cyber

to kinematics. Further scenarios are used throughout this book to illustrate ever
more advanced ISRT concepts.

Endnotes
[1] Feng, W., Y. Li, X. Yang, Z. Yan, and L. Chen, “Blockchain-Based Data Transmission
Control for Tactical Data Link,” Digital Communications and Networks, 2020.
[2] Eyer, K., and S. McJessy, “Operationalizing Distributed Maritime Operations,” March 5,
2019, https://cimsec.org/operationalizing-distributed-maritime-operations/.
[3] Richardson, J. M., “A Design for Maintaining Maritime Superiority,” Naval War College
Review, Vol. 69, 2016.
[4] Popa, C. H., et al., “Distributed Maritime Operations and Unmanned Systems Tactical
Employment,” Monterey, CA, 2018.
[5] Mattis, J., “Summary of the 2018 National Defense Strategy of the United States of Amer-
ica,” 2018.
[6] Larter, D. B., “Jim Mattis’ ‘Dynamic Force Employment’ Concept Just Got Real for the US
Navy,” Defense News, July 16, 2018, https://www.defensenews.com/naval/2018/07/16/
jim-mattis-dynamic-force-employment-just-got-real-for-the-us-navy/.
[7] Article 1(a) of the IMO Convention. For the IMO Focus on security, see https://www.
imo.org/en/OurWork/Security/Pages/Default.aspx.
[8] https://mc.nato.int/.
[9] https://www.nato.int/cps/en/natohq/topics_88745.htm.
[10] https://www.dst.defence.gov.au/.
[11] “National Plan to Achieve Maritime Domain Awareness, The National Strategy for
Maritime Security,” U.S. Coast Guard, 2005.
[12] Nimmich, J. L., and D. A. Goward, “Maritime Domain Awareness: The Key to Maritime
Security,” International Law Studies, Vol. 83, 2007.
[13] Bueger, C., and T. Edmunds, “Beyond Seablindness: A New Agenda for Maritime Security
Studies,” International Affairs, Vol. 93, No. 6, 2017.
[14] U.S. Government, “Joint Publication 2-0 Joint Intelligence,” 2013, Figure I-3. The
Intelligence Process, p. I-6.
[15] Brown, J., “Strategy for Intelligence, Surveillance, and Reconnaissance,” Air University
Press, 2014.
[16] Smagh, N. S., “Intelligence, Surveillance, and Reconnaissance Design for Great Power
Competition, Congressional Research Service,” Congressional Research Service, 2020.
[17] Hoehn, J. R., and N. S. Smagh, “Intelligence, Surveillance, and Reconnaissance Design
for Great Power Competition,” Congressional Research Service, June 4, 2020, p. 17,
https://crsreports.congress.gov/.
52 Naval ISR Fusion Principles, Operations, and Technologies

[18] “The U.S. Army in Multi-Domain Operations 2028,” 2018.

[19] “DoD Data Strategy,” 2020.
[20] “National Maritime Domain Awareness Plan for the National Strategy for Maritime
Security,” U.S. White House Office, 2013.
[21] Bommakanti, K., “India and China’s Space and Naval Capabilities: A Comparative
Analysis,” ORF Occasional Paper, 2018.
[22] Office of the Chief of Naval Operations, “The Navy Unmanned Undersea Vehicle (UUV)
Master Plan,” Technical Report, ADA511748, 11-09-2004.
[23] Button, R. W., et al., “A Survey of Missions for Unmanned Undersea Vehicles,” Santa
Monica, CA: Rand National Defense Research Inst., 2009.
[24] Chief of Naval Operations. U.S. Navy Program Guide 2017.
[25] Holler, R. A., A. W. Horbach, and J. F. McEachern, “The Ears of Air ASW – A History
of U.S. Navy Sonobuoys,” Warminster, PA: Navmar Applied Sciences Corporation,
2008, https://www.worldcat.org/title/ears-of-air-asw-a-history-of-us-navy-sonobuoys/
oclc/720627294.
[26] DARPA Ocean of Things, https://oceanofthings.darpa.mil.
[27] https://news.lockheedmartin.com/2006-01-18-Lockheed-Martin-Awarded-144-3-
Million-for-the-Navys-Advanced-Deployable-System.
[28] Board, N. S., and N. R. Council, “C4ISR for Future Naval Strike Groups,” 2006.
[29] Chief of Naval Operations. U.S. Navy Program Guide 2019.
[30] Board, N. S., and N. R. Council, “C4ISR for Future Naval Strike Groups,” 2006.
[31] https://defense.info/defense-systems/maritime-autonomous-systems-their-potential-
contribution-to-the-isr-mission-set/.
[32] Best, Richard A., C. C. Bolkcom, and Foreign Affairs, Defense, and Trade Division.
“Airborne Intelligence, Surveillance & Reconnaissance (ISR): The U-2 Aircraft and Global
Hawk UAV Programs,” Congressional Research Service, Library of Congress, 2000.
[33] https://www.navy.mil/Resources/Fact-Files/.
[34] “Space Operations,” Joint Publication 3-14, 2018.
[35] “Challenges to Security in Space,” Defense Intelligence Agency.
[36] Spires, D. N., “Beyond Horizons: A Half Century of Air Force Space Leadership,”
Colorado Springs: Air Force Space Command in Association with Air University Press, 1998,
p. 7.
[37] “Satellites,” Glactics, 1997, http://satellites.spacesim.org/english/main.html.
[38] Skinner, B., “Ground-Based Weapons: Kinetic Antisatellite Weapons,” Space Security
Index, 2020.
[39] Lervolino, P., R. Guida, P. Lumsdon, J. Janoth, M. Clift, A. Minchella, and P. Bianco, “Ship
Detection in SAR Imagery: A Comparison Study,” Proceedings of the IEEE International
Geoscience and Remote Sensing Symposium (IGARSS), July 23–28, 2017, pp. 2050–2053.
 Principles of Operations from Seabed to Space 53

[40] Van Etten, A., “Object Detection in Satellite Imagery, a Low Overhead Approach, Part
I,” August 30, 2016, https://medium.com/the-downlinq/object-detection-in-satellite-
imagery-a-low-overhead-approach-part-i-cbd96154a1b7.
[41] Millhouse, P., “Next-Gen Persistent Maritime Security with Commercial Space Assets,”
WEST 2020, March 2, 2020.
[42] The AIS is a maritime ship identification and location system that was mandated in
December 2004 by the IMO. All commercial vessels that travel in international waters
are required to transmit ship identification and location (derived from GPS) that can
be received by AIS receivers ashore, on other ships, or from satellites. AIS transponders
broadcast messages at regular intervals over two designated VHF channels.
[43] HawkEye typical geolocation performance for a trio of satellites varies between 0.1 to 0.2
km depending on satellite-to-target geometry at the time of measurement. The geolocation
is provided by: (1) Measuring and comparing time-of-arrival (TOA) and frequency-of-
arrival (FOA) of a signal between 2 or all 3 satellite receivers, (2) using GPS to maintain
precise estimates for the position and velocity of the spacecraft and their receive, and (3)
performing multilateration to estimate the signal emitter location. See Sarda, K., N. Roth,
R. E. Zee, Dan CaJacob, Nathan G. Orr, “Making the Invisible Visible: Precision RF-
Emitter Geolocation from Space by the HawkEye 360 Pathfinder Mission,” Proc. 32nd
Annual AIAA/USU Conference on Small Satellites, Paper SSC18-II-06, 2018.
[44] https://www.asc-csa.gc.ca/eng/satellites/radarsat/technical-features/radarsat-comparison.
asp.
[45] https://www.iceye.com/hubfs/Downloadables/ICEYE_SAR_Product_Guide_2021_
V4.0.pdf.
[46] https://www.newspace.im/constellations/planet-skybox.
[47] https://www.satimagingcorp.com/satellite-sensors/other-satellite-sensors/dove-3m/.
[48] “The National Military Strategy for Cyberspace Operations,” U.S. Government, 2006.
[49] Thiele, R. D., “Game Changer – Cyber Security in the Naval Domain,” ISPSW Strategy
Series: Focus on Defense and International Security, No. 530, 2018.
[50] Based on Joint Chiefs of Staff, Joint Terminology for Cyberspace Operations Memorandum,
Attachment 1 Cyberspace Operations Lexicon, November 2010.
[51] Joint Publication 3-32 Joint Maritime Operations, June 8, 2018, validated on
December 16, 2020.
[52] In naval tactics, a strategically placed island can be considered to be an unsinkable carrier
that provides maritime power projection from a hardened location. See Hughes, W. P., Jr.,
and R. P. Girrier, Fleet Tactics and Naval Operations, Third Edition, U.S. Naval Institute
Press, 2018.
[53] MSO are operations to protect maritime sovereignty and resources and to counter
maritime-related terrorism, weapons proliferation, transnational crime, piracy,
environmental destruction, and illegal seaborne migration.
[54] “Expeditionary Advanced Base Operations (EABO) Handbook Considerations for Force
Development and Employment,” Marine Corps Warfighting Lab, Concepts & Plans
Division, June 1, 2018.
54 Naval ISR Fusion Principles, Operations, and Technologies

[55] Cruise ships can traverse 10 km (at about 30 knots) in 10 minutes; it is reasonable for the
Blue fleet to traverse 100 km within sever­al hours to arrive on station. ­
3
Distributed Maritime Operations
The concept of distributing naval forces is relatively new because the fleet has
traditionally spaced surface assets to provide massed defense while enabling pre-
cisely coordinated attacks. The entire subject of naval tactics (the techniques
of action) is focused on protecting the fleet while enabling it to effect decisive
strikes. Esteemed naval tacticians, Captain Wayne Hughes (USN Ret.) and
coauthor Rear Admiral (RADM) Robert Girrier (USN Ret.) have cautioned:
“Today, missiles of various ranges and homing characteristics can be placed in
manned and unmanned ships and aircraft at a relatively low cost. With the
onset of information warfare concepts, this looks like the beginning of a lasting
change in naval warfare” [1].
Indeed, DMO introduces the wide distribution of forces, the arming of
every vessel, the introduction of weapons on unmanned vessels (surface and
subsurface) and aircraft, and a lasting change in naval warfare. With new na-
val concepts come new tactics, technologies, and operations that offer oppor-
tunities and risks that must be addressed. DMO introduces a drastic shift in
operations and tactics, and an increased dependence on technologies: sensors,
networks, computation, and decision support.
In this chapter, we introduce the tactical considerations for distributing
operations and the implications for C2ISR architecture and battle management.
In subsequent chapters we detail the technology considerations to provide the
ISR capabilities necessary for DMO.

55
56 Naval ISR Fusion Principles, Operations, and Technologies

3.1 Introduction to DMO

Since the Battle of Midway demonstrated the potential for airpower to domi-
nate naval engagements, the U.S. Navy structure has revolved around the car-
rier battle group. The potential for force projection by the carrier’s strike aircraft
constituted a dominant threat for decades when the United States control of
the seas and of the air was uncontested. For most of the 1990s and the early
2000s the USN force structure was based around the core of 12 carrier battle
groups [2]. During the same timeframe, U.S. naval strategy included having the
capability to field two amphibious warfare brigades standing ready to project
power ashore. This amphibious capability required 38 amphibious ships, which
is a number that was rarely met in practice. Most often, amphibious power was
deployed as marine expeditionary units (MEU) embarked on three large vessels
(landing helicopter dock (LHD) or landing helicopter assault (LHA), landing
platform dock (LPD), and landing ship dock (LSD)). However, this number is
being substantially reduced to 24 to 28 large amphibious ships in the FY2022
shipbuilding plan [3]. The reduction in large combatant vessels is, in part, due
to recognition that in recent years the threats posed by land-based, antiship mis-
siles (ASM) and enemy submarines have increasingly put such high value assets
at risk. As a result, a significant portion of the battle group’s offensive potential
is spent protecting the carrier rather than being available for strike operations
[4]. Indeed, as early as 1990, the Cato Institute called into question the level
of air and surface resources required to ensure a carrier’s survivability [5]. More
recently, an Air Force study challenged the notion that the carrier-based air-
wings could supply the overwhelming sustained strike capability that has often
been touted [6]. Concentrating combat power in a relatively small area has two
distinct disadvantages—simplifying the enemy’s targeting problem and limiting
the area over which our naval forces can project offensive capability. The CNO
issued a challenge in January 2016 to consider alternative fleet structures [7].
The concept that has now gained the most traction is DMO, which is devel-
oping as a means to address these two issues. The Navy’s current 30-year ship-
building plan represents a “once in a generation change” that pivots to support
the DMO concept [8]. By 2018, the CNO’s design for maintaining maritime
superiority 2.0 fully embraced DMO and directed continued development of
the concept [9].
DMO disperses combat power in ways that complicate the adversary’s
ability to engage friendly forces. Deploying in a larger number of smaller groups
of combatants and adding unmanned surface and subsurface vehicles as de-
coys causes the enemy to dilute surveillance—a tactic known as countertar-
geting. DMO goes beyond merely dispersing combatant vessels and integrates
expeditionary littoral bases from which land-based antiship and antiair mis-
siles become part of the network performing antiaccess and area denial (A2AD)
 Distributed Maritime Operations 57

functions. A2AD is a counterengagement strategy that relies on the ability to

threaten the enemy’s freedom of movement and limits their ability to concen-
trate force, making it difficult for the enemy to effectively mass combat power
sufficient to threaten our forces [10].
The offensive strategy of DMO has its roots in the concepts of distributed
lethality, which was widely discussed in the CNO’s guidance [9]. Rather than
concentrating combat power in a few highly capable ships, distributing lethal-
ity aims to go beyond the defensive considerations outline, and provide the
greater area covered by networked offensive weapons to hold enemy forces at
risk. Expeditionary bases can also be used to support short-term logistical needs
of surface, air, and subsurface assets to extend the range and duration of an op-
eration. The goal is not to control the seas everywhere at all times, but to have
the agility to hold any strategic maritime region at risk at any time.
The force structure and command and control systems needed to imple-
ment DMO is somewhat different than that which the Navy has maintained for
the last several decades. In particular, the distributed lethality necessary for ef-
fective DMO would imply a force structure that includes enhancing the lethal-
ity of individual minor combatants by increasing the ability of all combatants to
share intelligence and targeting information. All vessels can participate in lethal
action, even if organic sensors cannot reach targets; the fleet will network and
coordinate their sensors and weapons:

Enhanced surface ship lethality would allow smaller ships such as cruisers,
destroyers, and littoral combat ships to operate in hunter/killer SAGs that
could function autonomously, screen larger formations, and/or hold land
targets and sea lanes of communication/commerce at risk. To succeed,
it is imperative to distribute anti-air, ASW, and ant-surface fires. Future
weapons such as long range ASW weapons, rail guns, and new classes of
missiles may be needed to realize this level of lethality [11].

Networking for integrated fires requires persistent, organic over-the-

horizon (OTH), and thus overhead, surveillance. Broader deployment of the
Fire Scout unmanned aircraft system (UAS) is a first step, but integration with
airborne and particularly satellite surveillance will be critical to the success of
DMO. In order to process and disseminate intelligence and targeting informa-
tion to a dispersed network of vessels, a more robust and distributed command
and control paradigm must be adopted. Particular consideration needs to be
given to maintaining useability of the electromagnetic spectrum and to en-
suring secure cyber and communication protocols. While this book does not
provide details on methods to network assurance and cyber security, these are
implied capabilities that are covered in depth in texts dedicated to those sub-
jects. In the remainder of this chapter, we explore considerations for DMO
tactics and battle management in the DMO environment.
58 Naval ISR Fusion Principles, Operations, and Technologies

3.2 DMO Tactical Considerations

In his recent planning document, the CNO focuses attention on the necessity
of DMO by stating that “Emerging technologies have expanded the modern
fight at sea into all domains and made contested spaces more lethal. Ubiquitous
and persistent sensors, advanced battle networks, and weapons of increasing
range and speed have driven us to a more dispersed type of fight. Advances in
artificial intelligence and machine learning (AI/ML) have increased the impor-
tance of achieving decision superiority in combat” [12].
With such a force, where “everything is a shooter and a sensor” (or, as
RADM Fanta says, “if it floats, it fights” [13]) how does this impact the tactics
of naval forces? The force that is seamlessly interconnected by grids of networks
and is orchestrated with AI/ML–based decision support must impact fleet tac-
tics. Specifically, how does DMO influence tactics, or as Hughes contends, “…
the handling of forces in battle… where maritime issues are at stake” [14]?
DMO is defined by Braithwaite as “…an operations concept that leverag-
es the principles of distribution, integration, and maneuver to mass overwhelm-
ing combat power and effects at the time and place of our choosing…” [15]. In
the next paragraphs, we consider the implications of distribution, integration,
and maneuver, on naval tactics.
Distribution. DMO envisions distributed platforms, weapons systems,
and sensors across all domains. In contrast to the carrier battle group concept,
DMO seeks to “…‘distribute offensive capability geographically’: This speaks
to a wider dispersion of ships, in order to hold an enemy at risk from multiple
attack axes, and force that enemy to defend an increased number of vulnerabili-
ties, created by that dispersion. This point suggests what will become clear later,
and that is the disaggregation of forces, which is part and parcel of DMO.”
Thus, “a general paucity of assets in any high-end fight, in any theater can only
be addressed by the precise delivering of only the exact right force to the exact
right place at the exact right time” [16].
An example of the benefits of such a dispersion of forces is suggested by
Jensen:

Distributed maritime operations take a different point of departure.

Rather than attack mainland command and intelligence assets in a crisis
with China (e.g., AirSea Battle) or threaten commercial shipping lanes
(e.g., offshore control), small, dispersed land and sea detachments threaten
the ability of Chinese forces to concentrate from within their anti-access/
area denial umbrella. These forces deny Chinese freedom of movement
along key sea and air lines communication. Distributed forces change the
adversary’s cost calculus and buy time for flexible deterrence options and
assembling a joint task force [17].
 Distributed Maritime Operations 59

Integration. The confluence of vision and technological innovation have

driven the evolution of naval forces from collections of independent platforms,
weapons systems, and sensors to synergistic, integrated, systems-of-systems
sharing real-time information, affording every netted platform a common un-
derstanding of the battlespace with the same fire-control quality information.
Commanders and autonomous platforms share a common operational picture
and share a common fire-control solution to engage an adversary. Rear Admiral
Rodney Rempt projected this evolution to an even more sophisticated future:

A future in which the Navy’s tactical grid would one day be understood
as, simply put, an agnostic network of weapons and sensors, controllable
by any number of nodes, and without regard to where those weapons or
sensors or controlling nodes might be deployed or even in which unit
they existed. In the future, if an inbound threat were to be detected, this
agnostic, dispersed grid would determine which sensor(s) would be most
appropriate, and then, when necessary, the system would pair the most
capable and best located weapon with that sensor(s) in order to efficiently
engage the threat [16].

This sophisticated future envisioned by Rempt is a necessary tactical ad-

vantage for a victorious naval force. Rapid, decisive action, with an informed
understanding of the battlespace leads to maritime superiority.
Maneuver. Maneuver suggests the employment of rapid, unexpected, de-
terrent actions upon an adversary, allowing the exploitation of uncertainty and
achieving surprise. As a result, the adversary “…perceives he has lost control
and becomes the victim of disruption, confusion, and disorganization… The
adversary finds this rapid chain of unexpected events impossible to cope with
effectively” [18].
Distributed, highly integrated, maneuverable DMO-enabled forces, com-
prised of crewed and uncrewed vessels, are optimized for adaptability, which
leads to tactical advantage. Some examples of defensive and offensive engage-
ments yield further insight.
Distributed defensive strategies consider two distinct alternatives to a
multiaxis, multimissile engagement [19]:

• Clustering ships tightly to concentrate and distribute defensive capabili-

ties to avoid any one ship being overwhelmed;
• Dispersing ships widely (DMO) while sharing adversary warning and
targeting data to enable the distributed assets to respond cooperatively
in defense, as a single fighting unit.

In considering offensive action or coordinated attack:

60 Naval ISR Fusion Principles, Operations, and Technologies

…The offence might use a mix of crewed and uncrewed vessels. One
option is to use three ship types: a large, well-defended crewed ship that
carries considerable numbers of various types of long-range missiles but
which remains remote to the high-threat areas; a smaller crewed warship
pushed forward into the area where adversary ships are believed to be both
for reconnaissance and to provide targeting for the larger ship’s long-range
missiles; and an uncrewed stealthy ship operating still further forward in the
highest risk area primarily collecting crucial time-sensitive intelligence and
passing this back through the smaller crewed warship onto the larger ship
in the rear… An alternative concept is to have a large crewed vessel at the
center of a networked constellation of small and medium-sized uncrewed
air, surface and subsurface systems. A large ship offers potential advantages
in being able to incorporate advanced power generation to support
emerging defensive systems like high energy lasers or rail guns. In this, the
large crewed ship would need good survivability features, suitable defensive
systems, an excellent command and control system to operate its multitude
of diverse uncrewed systems and a high bandwidth communication system
linking back to shore-based facilities and data storage services [19].

The integration of distributed platforms, weapons systems, and sensors

has profound implications on tactical advantage. This advantage is manifested
in the synergies afforded by distribution, integration, and maneuver. Together
these attributes of DMO enable enhanced maritime situational awareness, real-
time sharing of a common operational picture and fire-control quality infor-
mation across all platforms, and ultimately a decision advantage that enables
action well before an adversary can react.
These factors—distribution, integration, and maneuver—are literally
game changers, as noted by Williamson, “An observable ocean will significantly
change naval operations, and today’s game of Battleship will give way tomorrow
to chess when every player can see the gameboard clearly. In that case, the ad-
vantage will go to the strategist who can best foresee future moves, understand
environmental factors, and take the most rapid, decisive actions to achieve the
advantageous position most consistently” [20].

3.3 DMO Architecture and Elements

In formulating an architecture capable of achieving the connectivity needed for
distributed networked fires in real time, it is useful to consider a tactical grid
of connected, distributed nodes. The nodes can be sensors, weapons, commu-
nications, and decision nodes. Historically we have seen this concept mature
through several phases. As missile technology began to surpass direct fire guns
in the 1950s it became necessary to share targeting information between mis-
sile-capable vessels. The result was the tactical data line (TDL, or TADIL) sys-
 Distributed Maritime Operations 61

tems. The next major evolution was the Aegis system, which matured through
the 1970s and allowed coordination of all of an Aegis-capable battle groups
weapon systems. Aegis originally was envisioned as a defensive system coordi-
nating disparate weapons systems such as the Phalanx® close in weapon system
(CIWS)—a 20-mm gatling gun for short-range missile defense—and the Mk-
41 vertical launch system (VLS) capable of deploying a variety of missiles for
self-defense and strike. The primary sensor for Aegis is the AN/SPY-1 radar
which is capable of search, track, and midcourse missile guidance functions but
other short-range sensors and terminal guidance radars are integrated as well.
The latest evolution, made publicly known in the early 1990s, is the CEC,
which allows for sensors to share raw data, as opposed to high-level symbolic
information about the situation. This allows for a combatant vessel to view
another vessel’s radar as its own and develop actual targeting solutions for its
weapon system based on the other vessel’s sensor feed [16].
DMO would take CEC a step further and would leverage AI to imple-
ment decision nodes that are tightly coupled and support high-level human
decision making in the presence of the chaos of battle. The AI could provide de-
cision support logic to determine which vessels were in the best position to fire
upon an enemy, what weapons they should employ, and which sensors should
support the engagement. At the same time, defensive decisions will also be
made, such as which vessels are at risk of attack and what countermeasures they
should employ. Georeferenced data, automated situation recognition, machine
learned detection of activities, knowledge of observed and models of antici-
pated adversary TTPs, may all be employed to recognize threatening behaviors
[21]. Using this prior knowledge and applying multisensor fusion analysis can
all be processed within the DMO context to provide decision support by AI at
speeds far greater than humans alone could accomplish. The resulting decision
superiority (timely and effective decision making) is a critical combat multiplier
in the DMO paradigm.
The various nodes to be integrated into the tactical grid include not only
the traditional set of sensors and weapons that already exist in the fleet, but
also the emerging technologies, platforms, and weapon systems that will enable
DMO [22]. This could include USVs of various types, such as large USVs that
serve as weapons platforms, perhaps deploying the proposed long-range anti-
ship missile (LRASM), and medium-sized USVs that would function primarily
as sensor platforms.
Unmanned air platforms will also be key nodes in DMO. The MQ-8B
Firehawk can be deployed from a variety of ships since it requires only a small
helipad to launch. Having trained operators and maintenance personnel is still
necessary, however. The MQ-8B Firehawk can serve as both a sensor and a weap-
ons system. Firehawk is equipped with an RDR-1700B X-band radar capable
of all-weather search and targeting and can share that targeting information
62 Naval ISR Fusion Principles, Operations, and Technologies

with other platforms via CEC. Should immediate engagement be required,

Firehawk is equipped with the advanced precision kill system (APKS)—laser
guided 2.75-in rockets [23]. The Scan Eagle UAS provides a different kind of
flexibility with pneumatic launch and skyhook recovery and provides persistent
day/night and thermal imagery for over 20 hours on station [24].
In order to ensure reliable, high bandwidth communications OTH, it will
be necessary for the DMO architecture to incorporate communications relay
nodes. There are two basic options for such nodes: satellites or UAS. While
geostationary communication relay satellites have been used in the past, the
need for high bandwidth and low latency are driving the solutions to LEO sat-
ellites [25]. The commercial world is already addressing high bandwidth satel-
lite infrastructure with products like the Starlink satellite internet constellation,
offering 60 to 90 Mb/s rates or higher for residential customers. Whether the
military will procure bandwidth from such commercial providers or invest in
their own technology has been hotly debated over the last few years [26]. The
trade-offs are lower cost for commercial systems versus direct control, secure
protocols, and availability of military systems. Some vendors are marketing di-
rectly to DoD as a result. A notable example is the Viasat XVI, a Link-16 en-
abled satellite communication relay [27]. There are many UAS options which
may offer low-cost alternatives [28], ranging from dedicated Global Hawk relay
platforms to the small, unmanned multirotor aerial relay [29].

3.4 All-Domain C2 Battle Management

At present, overall maritime mission command and control falls under the Joint
Maritime Combatant commander. Responsibility is then further distributed to
separate commanders: first to regional task forces (as the force is usually dis-
tributed geographically), and then by mission and domain. Each task force will
have a commander for each of the following roles for which their task force has
capability and responsibility:

• Surface warfare;
• Air and missile defense;
• Antisubmarine warfare;
• Mine warfare;
• Strike warfare;
• Amphibious operations.

There are a host of other specialized minor operations requiring delin-

eation of C2 requirements as well, ranging from counterdrug and maritime
 Distributed Maritime Operations 63

interdiction to noncombatant evacuation, humanitarian support, and logistical

missions such as maritime prepositioning [30].
The location of the C2 centers for these operations has traditionally been
selected based on available communications, cryptographic, and sensor resourc-
es, as well as proximity to the operational assets involved and the geographical
considerations of the mission itself. A major limitation on DMO is that al-
though one can assert that “if it floats, it fights” [31], that different ship types
are necessarily optimized for certain roles and staffed with sailors and com-
manders with expertise in those roles. Designating a mine-sweeping ship as the
antiair mission commander would never be ideal, for instance. In such a case,
none of the required sensors, weapons, or human capital would be resident on
the minesweeper, and would all need to be imported from other nodes, impos-
ing latency and reliability concerns. It is doubtful that DMO can be extended
so far as to view specialized ships as homogeneous assets. Nonetheless, the abil-
ity to share the totality of the operational picture among several ships provides
greater resilience by enabling the seamless transfer of operational command
from one ship to another, or even to shore.
A useful principle for deciding where operational command should physi-
cally reside is to consider the observations of former Pacific Fleet commander,
Admiral Scott Swift. Swift has pointed out that a century ago, naval officers un-
derstood that C2 was about leadership and tactical expertise. Over the course of
the late twentieth century, communications and computers have been included
and the term C4 has replaced C2. Swift thinks this is a mistake because it con-
flates the technical means of achieving control with the cognitive processes of
making command decisions [32]. Regardless of whether electronic communi-
cation and decision support tools are available, a ship’s captain must still com-
mand his ship. For this reason, Admiral Swift published a set of fighting orders
in part to serve as guidance to commanders who have been cut off from their
higher C2 elements during distributed operations [16]. DMO operations are,
of course, dependent on reliable and secure networking to assure the benefits of
distributed and highly automated decision making, but commanders must be
prepared to degrade gracefully (or abruptly) and sustain operations.
Distributing classified information will be required to achieve informa-
tion sharing at the level required for DMO. In addition to requiring appropri-
ate cryptographic assets and classified information systems on each ship, the
effort to secure classified data in motion will increase substantially. More classi-
fied data will be exchanged over radio frequencies, meaning more transmissions
over longer ranges, which increases exposure to interception.
Sharing the common operational picture is currently done in the U.S.
Navy to provide high-level situational awareness. However, to enable DMO
requires sharing actionable tactical information which would support target-
ing decisions, weapons selection, and more. This capability has been described
64 Naval ISR Fusion Principles, Operations, and Technologies

as a battle force manager (BFM) [16]. The entire BFM functionality would
need to be replicated on each ship that would be capable of assuming the mari-
time operational commander (MOC) role. BFM capabilities would include at
a minimum:

• Maintaining connectivity to all nodes;

• Monitoring the status of the ordinance on every node;
• Health and status of each node, such as damage and fuel;
• Understand operational capabilities and limitations of each node;
• Ability to change operational parameters in any sensor or weapon sys-
tem;
• Explainable AI-based decision support that describes distributed, auto-
mated decision recommendations, rationale to allow rapid concurrence,
adjustment, or rejection by decisionmakers in the CIC;
• Explainable sensor/weapon/target pairing algorithms and recommenda-
tions;
• Data storage, sorting, analysis, and presentation.

3.5 A DMO Scenario

Building on the scenario introduced in Chapter 2 we provide a particularly illu-
minating vignette based on a published challenge problem by Eyer and McJessy
[16] to portray the disruptive benefits of DMO. This scenario details the inte-
grated response of a DMO-enabled naval system of systems as it confronts a so-
phisticated nation-state threat with hypersonic weapons, and one can postulate
that the DMO-enabled force includes assets from seabed to space.
This scenario (Figure 3.1) takes place in the Kandago Sea where two na-
tions have contested claims over a series of ten islands across a 400-km chain
in the Kandigan Straits; as in Chapter 2, the Red naval forces are beginning an
assault on the contested island chain and distributed Blue forces, 350-km out to
sea, are moving to defend the islands in support of Green nation.

1. In this case, Red nation will attack the approaching Blue naval fleet
with hypersonic glide missiles from two ground batteries located on
the coast. The flight time of these Mach 7 missiles is less than 5 min-
utes from coastal launch to the Blue fleet 350 km away.
2. In this agnostic grid, the launch is detected by multiple, mutually re-
inforcing methods, including (1) a constellation of IR sensing satel-
lites and (2) a constellation of relay satellites that relay the IR signature
 Distributed Maritime Operations
65

Figure 3.1 DMO configuration in the Kandigan Straits.

66 Naval ISR Fusion Principles, Operations, and Technologies

of missiles to the fleet and to other satellites to enable rapid missile ac-
quisition, handoff between satellites, and enable the fleet to establish
a track. The sensing satellites coordinate a combination of wide and
medium field-of-view sensors to detect, track, and then handoff track
custody to close-in defensive weapon systems in the Blue fleet [33].
3. The Blue fleet shipboard phased array radars are automatically changed
to maximize its tracking capability. As more sensors are automatically
brought to bear, a precise track, including origin and aim point, is
generated.
4. At the same time, decisions are made at the strategic and operation-
al levels, decisions dramatically aided by the application of artificial
intelligence: Is the threat real? What asset(s) are under threat? What
hard- and soft-kill techniques and systems are best employed? What
systems are both in position and possess the capability and capacity
necessary for engagement? What is the optimal engagement timeline?
What additional sensors should be brought to bear, and when? Jam-
ming? Chaff? Decoys? From whom and when? Who shoots? When do
they shoot? What ordinance do they shoot? How many rounds?
5. Orders are automatically issued to concerned units, yet the entire
network, including other decision nodes, remains fully cognizant of
the larger picture. The battle management system has built-in redun-
dancies so that if one node is destroyed, another automatically and
seamlessly steps in. All of these decisions can be automated, if desired,
in order to maximize speed and the optimal response, provided that
commanders allow for that automation.

Ultimately, only the necessary and best systems are matched to the threat,
at only the right time, maximizing effect and minimizing the waste of lim-
ited resources. The most effective and efficient method of engagement becomes
routine.
Note the diversity of sensors implied, in type and capability, their disper-
sion, their seamless integration, and the orchestration provided by AI-enabled
decision support agents that can reason on timelines appropriate for the find,
fix, and finish process necessary to defeat a sophisticated threat. Also of note, is
the resilience that is offered by DMO-enabled forces, in particular their seam-
less integration. In the event one node is destroyed, other nodes with an identi-
cal operational picture and firing solution readily step in.
 Distributed Maritime Operations 67

Endnotes
[1] Hughes, W. and R. Girrier, Fleet Tactics and Naval Operations, Annapolis, MD: Naval
Institute Press, 2018, p. 132.
[2] “U.S. Ship Force Levels,” Naval History and Heritage Command, https://www.history.
navy.mil/research/histories/ship-histories/us-ship-force-levels.html
[3] Cancian, M., “Stormy Waters Ahead for Amphibious Shipbuilding Plan,” July 1, 2021, https://
breakingdefense.com/2021/07/stormy-waters-ahead-for-amphibious-shipbuilding-plan/.
[4] Torvold, W. D., Defending the Aircraft Carrier: Doctrine and Technology for Survival. New-
port, RI: Naval War College, 2000.
[5] Isenberg, D., “Cato Institute Policy Analysis No. 134: The Illusion of Power: Aircraft Car-
riers and US Military Strategy,” 1990.
[6] Grant, R., “The Carrier Myth,” Air Force Magazine, Vol. 82, Issue 3, 1999.
[7] Richardson, J. M., “A Design for Maintaining Maritime Superiority. Version 1.0.,” Chief
of Naval Operations Washington United States, 2016.
[8] O’Rourke, R., “Navy Force Structure and Shipbuilding Plans: Issues for Congress,” Con-
gressional Research Service Report 19, 2018.
[9] Richardson, J. M., “A Design for Maintaining Maritime Superiority,” Washington D.C.:
U.S. Navy, 2018.
[10] Jensen, B., “Distributed Maritime Operations: Back to the Future?” April 9, 2015, https://
warontherocks.com/2015/04/distributed-maritime-operations-an-emerging-paradigm/.
[11] Rowden, T., P. Gumataotao, and P. Fanta, “Distributed Lethality,” Proceedings Magazine
Vol. 141, No. 1, 2015, 1343.
[12] Gilday, M., CNO NAVPLAN, Washington, D.C.: United States Navy, 2021.
[13] Freedberg, S. J., “‘If It Floats, It Fights’: Navy Seeks ‘Distributed Lethality’,” Retrieved from
Breaking Defense, January 14, 2015: https://breakingdefense.com/2015/01/if-it-floats-it-
fights-navy-seeks-distributed-lethality/.
[14] Hughes, W. P., Fleet Tactics and Coastal Combat, Annapolis, MD: Naval Institute Press,
1999.
[15] Braithwaite, K. J., “Advantage at Sea Prevailing with Integrated All-Domain Naval Power,”
Washington D.C.: Department of the Navy, 2020.
[16] Eyer, K., and S. McJessy, “Operationalizing Distributed Maritime Operations,” Center
for International Maritime Security, 2019, http://cimsec.org/operationalizing-distributed-
maritime-operations/39831.
[17] Jensen, B., “Distributed Maritime Operations: Back to the Future?,” War on the
Rocks, 2015, https://warontherocks.com/2015/04/distributed-maritime-operations-an-
emerging-paradigm/.
[18] Wilson, G. I., “The ‘Manuever Warfare’ Concept,” Gazette Magazine, Marine Corps
Association, July 17, 2019, https://mca-marines.org/blog/gazette/the-maneuver-warfare-
concept/.
68 Naval ISR Fusion Principles, Operations, and Technologies

[19] Layton, P., “Fighting Artificial Intelligence Battles: Operational Concepts for Future AI-
Enabled Wars,” 2021.
[20] Williamson, W., “From Battleship to Chess,” U.S. Naval Institute Proceedings, July 2020,
Vol. 146, No. 7, https://www.usni.org/magazines/proceedings/2020/july/battleship-chess.
[21] TTPs refer to three categories of adversary behaviors:
Tactics: The employment and ordered arrangement of forces in relation to each other;
Techniques: Nonprescriptive ways or methods used to perform missions, functions, or
tasks;
Procedures: Standard, detailed steps that prescribe how to perform specific tasks.
[22] 2019 Surface Navy Symposium in Washington D.C.
[23] “Fire Scout Unmanned Aircraft System,” https://www.northropgrumman.com/what-we-
do/air/fire-scout/.
[24] ScanEagle, https://www.boeing.com/defense/autonomous-systems/scaneagle/index.page.
[25] Erwin, S., “Space Force Finalizing Plan to Procure Broadband from Low-Orbit
Satellites,” March 27, 2021, https://spacenews.com/space-force-finalizing-plan-to-
procure-broadband-from-low-orbit-satellites/.
[26] Hitchens, T., “Griffin: DoD Can’t Rely on Commercial Satellite Communications,”
December 3, 2019, https://breakingdefense.com/2019/12/griffin-dod-cant-rely-on-
commercial-satellite-communications/.
[27] Miller, A., “Powered by Satellite, Link 16 Military Radio Set to Get a Huge Boost,”
June 27, 2019, https://www.viasat.com/about/newsroom/blog/xvi/.
[28] Carr, D. F., “Communications Relay Grows with Expansion of UAV Missions,”
August 3, 2009, https://defensesystems.com/articles/2009/07/29/c4isr-1-uav-relay.aspx.
[29] “Unmanned Multirotor Aerial Relay (UMAR),” https://www.dragonflypictures.com/
products/umar-tethered-uas/.
[30] Joint Publication 3-32, Joint Maritime Operations, June 8, 2018.
[31] Rowden, T. S., “Surface Force Strategy: Return to Sea Control,” San Diego, CA: Naval
Surface Force Pacific Fleet, 2016.
[32] Swift, S., “Master the Art of Command and Control,” Proceedings of the U.S. Naval Institute,
Vol. 144/2/1,380, February 2018, https://www.usni.org/magazines/proceedings/2018/
february/master-art-command-and-control.
[33] This is based on the U.S. Space Development Agency constellation concept. See
Erwin, S., “The Pentagon’s Hyperfocus on Hypersonic Missile Threat,” Space News,
August 25, 2012, https://spacenews.com/the-pentagons-hyperfocus-on-hypersonic-
missile-threat/.
4
Naval Information Fusion Systems
Navies have traditionally partitioned sensor systems into stovepipes in the mis-
sion areas of air warfare, surface warfare, antisubmarine warfare, electronic war-
fare, and missile defense. Despite the criticisms of stovepipes, this approach
simplified acquisition of independent systems and enabled development and
procurement to be focused on mission effectiveness in each area. TDLs inter-
connected these systems to the Combat Information Center (CIC), and the
CIC to shooters (weapon systems) to engage targets. Integration of information
from each mission area occurred at the centralized CIC and a common operat-
ing picture could be distributed to all participants that had access via TDLs or
broader satellite communications.
This portioning made sense and has been efficient to perform any auto-
mated fusion of sensor data just within, but not across, these mission domains.
These functions merged their respective information about threats (objects,
tracks, events, and activities) in the CIC where the commander and staff ap-
plied the current context to understand the situation presented to them.
The desire to perform all-domain sensing brings the need to expand the
domain coverage, increase the speed of sensing and fusion, and cause ISR to de-
liver targeting-quality information is a new factor driving significantly expand-
ed automation. We now use the term ISRT to refer to this close integration of
ISR to C2 and targeting. The desire for increased networking, NCW, and joint
all-domain ISR-C2, requires even greater integration of systems.
In this chapter, we examine a baseline of the current categories of naval
sensing and fusion systems and information integration to perform joint ISR
and C2, using, as an example, general U.S. naval systems that are representative
of advanced international systems. We also introduce the computing and net-
working concepts and technologies that are enablers to extend the integration

69
70 Naval ISR Fusion Principles, Operations, and Technologies

to a greater degree—to integrate organic and nonorganic sensors and sources,

the subsurface and the seabed.
While this chapter introduces the baseline state of the art, we will then
provide greater depth in Chapters 6 (the data fusion process) and 7 (the net-
work and distributed data fusion). The application of these technologies for
naval ISR is being developed by international navies as they seek to apply AI
capabilities [1].

4.1 Enterprise-Level Fusion

The naval information enterprise is expansive and includes many diverse el-
ements enumerated in Table 4.1, both afloat and ashore. It is a component
of a larger information enterprise that encompasses all other service elements
(Army, Air Force, Marines) that participate in all-domain operations, as well as
the forces of partner nations that operate in a coalition.
While many functions exist across these elements (e.g., cyber warfare, C2,
EW, battle management, oceanography, and meteorology), we focus on just
the many functional elements that comprise the ISR portions of the enterprise
(Table 4.2).
The integration of information across all these elements to perform dis-
tributed operations requires unprecedented coordination of standards, data rep-
resentations, interoperable data links, distributed computing resources, sharing
mechanisms and operating concepts, all at a fine level of granularity and at a
massive enterprise scale.
At this highest level of abstraction, the enterprise requires a common for-
mal understanding of fundamental terms, formats, and processes, for example:

• Enterprise governance (information technology (IT) strategy,

resource management, risk management, performance measurement,
optimization);

Table 4.1
Elements of the Naval Information Enterprise
Afloat Ashore
·· Task force command; ·· National command centers;
·· Strike group command(s); ·· Intelligence centers (national, services);
·· Organic sensors (seabed-to-air); ·· Naval fleet commands;
·· Organic TDLs (seabed-to-air); ·· Space sensing systems;
·· Space communication links. ·· Space communication, navigation centers;
·· Meteorological and oceanographic systems (METOC) centers.
 Naval Information Fusion Systems 71

Table 4.2
Major Categories of ISR and Information Operations (IO) Sensing,
Processing, and Exploitation Subsystems
Comparable USN
ISR Areas ISR Subsystems Capabilities Systems
ISR Enterprise Common computing environment with service-oriented Consolidated afloat
architecture to enable ISR and C2 services to share networks and
information, maintain a common understanding of the enterprise devices
battlespace, and coordinate sensing across all services. (CANES), integrated
shipboard network
system (ISNS)
Maritime domain ISR aggregating, correlating, and fusing all source Defense common,
awareness intelligence in real-time and near real-time processing ground system
(combat analysis for tracking, targeting, and exploitation. DCGS-N
capability all Coordination of both afloat and ashore elements.
source data)
Maritime domain Visualize vessel positions and tracks on a map SeaVision data,
awareness (terrestrial and satellite AIS, coastal radar, and satellite fusion service
(unclassified imagery based positions in time (POSITS)). Complex
for commercial searches, alerting, and vessel risk scoring. Multiple
vessels) source correlation.
Air surveillance Detect and visualize air tracks (commercial, combat) AEGIS Combat
and air defense from all sources (radar, IFF, and AIS) and identify tracks; System; surface
designate for challenges and engagement. search radar systems
Detect and track surface vessels and identify tracks;
designate tracks for engagement.
Missile defense Detect and track ballistic missiles near midcourse (post- AEGIS Ballistic
boost and pre-reentry) to relay track data to land-based Missile Defense
interceptors or to engage with ship based terminal System
phase interceptor missiles.
Signals Signals intelligence sensing and emitter location Ships signal
intelligence/ and identification, apply electronic disruption, denial, exploitation
electronic exploitation, deception, and attack mechanisms. equipment (SSEE)
warfare/ Monitor adversary IO activities (MILCOMM, broadcast,
information social media, etc.) and issue approved IO responses.
operations
Antisubmarine Perform surveillance, search, detection, track, and Undersea warfare-
warfare (ASW) localization of submarines for attack. Systems can be decision support
and seabed deployed on or from individual platforms (surface, air, system (USW-DSS)
warfare underwater, space) and ASW operations are carried out
by individuals or coordinated groups of platforms.
Cyber and Protect cyberspace and the electromagnetic spectrum CANES SIEM
network defense from adversary attack, detect attacks, locate threat(security information
vectors, respond, and mitigate effects. and event
management)
Meteorological Acquire, analyze, and predict the elements of the Global command and
and physical environment that affect naval warfare, acquire, control system-
oceanographic analyze, and predict the elements of the physical maritime (GCCS-M)
systems (METOC) environment that affect naval warfare.
72 Naval ISR Fusion Principles, Operations, and Technologies

• Core functions used across the enterprise: (e.g., content management

(storage, search, discover, archive), multilevel security, cyber security, vi-
sualization);
• Data standards for:
• Tagging data (source, time, modes, target ID assigned, etc.);
• Tagging security (source security level and aggregated or fused data
security level) for each maritime object of interest level;
• Metadata describing the pedigree of data;
• Discovery of services and data enterprise-wide.
• Security standards for network defense and cybersecurity;
• Messaging and collaboration;
• Standards and specifications by hardware and infrastructure providers
and software applications.

Even small components of functionality may be standardized as building

blocks that can be combined on the fly to support rapid-response operations
and the insertion of new capabilities.
The implementing process to modularize these functions to be shared
and reused is the service-oriented software architecture implemented in cloud-
based computing environments [2]. And the extension of service orientation
functionality over a distributed naval fleet afloat and it support elements ashore
requires a robust information network infrastructure. The next paragraphs in-
troduce these capabilities.

Information Environment
A robust and secure network infrastructure, from seabed-to-space, is required to
realize comprehensive ISR across distributed maritime assets. In the U.S. DoD
this is referred to as the global information grid (GIG), then the DoD informa-
tion network (DoDIN) and more recently has been referred to as a component
of DoD cyberspace in Joint Publication 3-12, Cyberspace Operations, June 8,
2018. The network is formally defined as:

The globally interconnected, end-to-end set of information capabilities for

collecting, processing, storing, disseminating, and managing information
on demand to warfighters, policy makers, and support personnel. The GIG
includes owned and leased communications and computing systems and
services, software (including applications), data, security services, other
associated services, and National Security Systems. Non-GIG IT includes
stand-alone, self-contained, or embedded IT that is not, and will not be,
connected to the enterprise network [3].
 Naval Information Fusion Systems 73

Like business enterprises, the military information enterprise, of which

the network infrastructure is core, includes all information resources, assets,
and processes required to achieve an information advantage and to share in-
formation across DoD and with mission partners; this enterprise includes [4]:

• The information itself (e.g., sensor data, processed data, fused intelli-
gence, sensor, and process management data);
• The processes associated with managing information to accomplish the
mission and functions (e.g., ISR sensor processing, sensor management,
fusion, visualization);
• Activities related to designing, building, populating, acquiring, manag-
ing, operating, protecting, and defending the information enterprise;
• Related information resources such as personnel, funds, equipment, and
IT, including internal use software and national security systems.

The layers of the information environment and the location of high-level

ISR elements (Figure 4.1) are like those in any large business enterprise; in the
business case ISR is like marketing and sales roles (marketplace awareness and
customer interaction-delivery). But the analogy is limited; business enterprise
operations are optimized for profit with acceptable loss factored in—a dollar
loss has the same value as a dollar gain. This is not true for naval operations in
general and certainly not for naval ISR where a missed or misidentified target
can be catastrophic. The vertical bars in the architecture refer to core services
(mission assurance and information management) that cross all layers (e.g., se-
curity services such as encryption for protecting data are common at all layers).

Service-Oriented Architecture
In modern navies, the evolution of such enterprise-level ISR systems has re-
sulted in the adoption of service-oriented architecture (SOA). With a SOA, the

Figure 4.1 Conceptual Layers of the IE and the ISR applications.

74 Naval ISR Fusion Principles, Operations, and Technologies

enterprise ISR workflows are performed by services; each service is identified,

its function is explicitly specified, and it can be discovered (computationally) by
other services and integrated dynamically into a processing flow. The services
create a data flow by asynchronous message (and data) passing across a distrib-
uted SOA computing system.
For ISR fusion, this means that the core fusion functions (e.g., GEOINT
object extraction, data alignment (in space, time), correlation, and association,
SIGINT-GEOINT fusion, or SIGINT extraction) can be assembled into dif-
ferent processing chains (or workflows) and the services tailored (parameter-
ized) for different fusion problems, yet the core services remain standard [5].
Structuring the overarching ISR enterprise in this manner results in several ben-
efits (Table 4.3).

Table 4.3
Main Features and Benefits of SOA
Features Characteristics and Benefits to ISR Processing Workflows
Service ·· Service provision provided by a federation of common resources;
·· Services enable improved information flow in distributed computing environment;
·· Ability to expose internal functionality;
·· Workflow flexibility as one process of an ISR workflow may be executed in
resource (Service A) while another process may be executed in a different one
(Service B).
Service reuse ·· Lower software development and management cost as common services are
reused by many workflows;
·· Commercial service providers act in an open market by advertising their services;
end users select providers based on their offerings.
Service ·· Ability to develop and integrate new ISR functional capabilities rapidly by
composition defining new service sequences using a set of exiting services.
Service ·· Ability to optimize performance, functionality, and cost by a simpler introduction
discovery of system upgrades (new services).
Service ·· Services communicate by message passing and are therefore loosely coupled
interaction (operating systems and programming languages are hidden to other services);
·· Loosely couple interaction provides flexibility, replaceability, scalability, risk
mitigation, and fault tolerance;
·· Web services provide a messaging between distributed software services using
a common platform and language independent standards (e.g., HTTP, Web
Services Definition Language (WSDL), Simple Object Access Protocol (SOAP), and
Extensible Markup Language (XML)) [6].
Asset ·· Services can integrate (wrap) existing code assets to apply legacy functionality.
wrapping
Message ·· Performance measurement measured by message efficiency;
monitoring ·· Security attack detection can be performed by message activity monitoring.
Message ·· Message management policy can control message throughput, routing;
control ·· Application of security policy can prevent message attacks and provide data
confidentiality and integrity.
Virtualization ·· Improved reliability by implementing services in virtual machines (VMs);
·· Ability to scale operations to meet different demand levels.
 Naval Information Fusion Systems 75

In the next section we introduce the information environment and illus-

trate the operation of an ISR workflow across services in a SOA environment,
describing the asynchronous message passing.

4.2 Information Systems Fusion

The distributed architecture provides responsive, and timely information be-
cause of the integration of interoperable services and information that is shared
across those services. The approach also allows navies to leverage significant
commercial industry IT investments in the technologies associated with SOA.
Of course, this ability to integrate (fuse) information contributes to the objec-
tive of achieving a decision advantage and enabling distributed maritime op-
erations and lethality. Commanders at all echelons require information that is
critical to their understand of the maritime situation and their options (courses
of action) for ISR and C2.
Table 4.4 provides examples of many of these questions for the Kandigan
Straits scenario in this book. These are based on typical intelligence preparation
in the maritime operating environment [7]. The questions are organized by
three of the levels of the JDL fusion model, introduced briefly in Chapter 1 and
described in greater detail in the Chapters 6 and 7.
To answer these questions in a comprehensive manner the information
architecture (distinguished from the IT architecture in the previous section) is
the structure that organizes, assembles, and labels content in an effective and
sustainable way to perform the maritime mission:

• Organize content: Establishes the ontology or taxonomies and hierar-

chies of information, such as sensor and source data types (e.g., sensor
measurements and human source text reporting);
• Assemble content: Defines the mechanisms to structure data in data
stores, to age the data (when to archive or dispose of data), to discover
and access the data;
• Label content: Provides the format for labelling data in storage and trans-
mission. This includes defining metadata standards for ISR data at the
source (sensor time-of-collect, sensor ID, sensor mode at time-of-col-
lect, sensor pointing, platform latitude-longitude, etc.).

Stovepipe ISR systems maintained an information architecture that ad-

dressed the functions enumerated in Table 4.2 for each domain stovepipe. An
integrated systems requires a common information architecture to allow infor-
76 Naval ISR Fusion Principles, Operations, and Technologies

Table 4.4
Key Commander’s Questions at Three JDL Levels
All-Domain Situation and
Warfare Object (Target) Domain Situation Threat Assessment (JDL L2
Domain Refinement (JDL L1) Assessment (JDL L2) and L3)
Space ·· Commercial and ·· Which Blue force Threat:
adversary overhead vessels may be ·· What access does Red have
tracks and next observable by Red? to space assets (commercial,
observations ·· What is the maritime allied partners, national
windows (EO, IR, RF); coverage of space technical capabilities) that
·· Blue critical Comm. assets by Blue? expose Blue operations?
and nav. space ·· Are Blue space assets What are the limitations of
assets; at-risk, and what would Red space access?
·· Adversary counter- be the impact? ·· What comms are red
space assets (land dependent on for space ISR?
and space) [8].
Air ·· Aircraft at Red ·· What is the Red air All-domain situation:
military airfields and force composition, ·· What is the Red COA?
use of commercial readiness, and ·· What is the timing of the
airfields; operations? COA and what are their
·· Red military aircraft ·· What are potential critical decision points, their
tracks; red air operations and vulnerabilities to achieving
·· Potential supporting abilities to support objectives?
air tracks. maritime COAs? ·· Does ISR support and
describe the range of
anticipated COAs?
All-domain threat:
·· Is there a coordinated
offensive?
·· When will they likely attack?
·· Where is the deception?
·· Which domain is at the
greatest risk?
·· Is the air and surface
maneuver coordinated for
a first-strike engagement
window?
·· Are the media and military
operations coordinated?
To what end? How can the
public media influence be
countered?
·· Why is the cyber-attack
occurring now? What is the
timeline?
·· What do all domains of
activities tell us about the
COA?
Human/ ·· Locate major ashore ·· Is Red exploiting media
social media influencers to influence Green
and sources of public public; how?
opinion. ·· What are the main
channels used?
 Naval Information Fusion Systems 77

Table 4.4 (continued)

All-Domain Situation and
Warfare Object (Target) Domain Situation Threat Assessment (JDL L2
Domain Refinement (JDL L1) Assessment (JDL L2) and L3)
Cyber ·· Detect and identify ·· Are cyber and EW
and EW EW activities and maneuver actions
maneuver sources (disruption, coordinated?
denial, deception); ·· What are the targets of
·· Cyber vectors and these actions, and likely
sources. next COAs?
Surface ·· Search, find, ID, and ·· Where is the missing
track all Red surface combatant vessel?
combatants and ·· How many vessels went
support vessels; dark?
·· Focus search (all ·· Have they deployed any
domains) for lost or SUVs? Where?
dark combatant or ·· Where is the adversary
potential-combatant missile range envelope?
vessels; Relative to current Blue
·· Search (surface, air) forces?
for small signature ·· Where are the
USVs conducting commercial vessel
reconnaissance or clusters that may
relay. confuse ISR?
Subsurface ·· Search, find, ID, ·· Why 2 subs to the
and track all Red north?
undersea vessels. ·· Is there a third sub
deployed? Where is it?
·· Have they deployed any
UUVs? Where?
Seabed ·· What are the ·· Are we in range of ·· What is the threat posed by
vulnerable seabed any known adversary seabed sensors; what is their
assets in the area of seabed sensors? range and detection ability?
ps (AO)? ·· What is the latency to ·· What if we have been
·· What are seabed Red seabed sensors? detected by seabed sensors;
threats, sonar nets, what countermeasures are
mines? available?

mation from all domains to be collected, stored, processed, and disseminated

effectively.
Consider the large architecture adopted by the U.S. Navy in transitioning
from ISR stovepipes to an integrated information and IT architecture [9]. The
architecture has the following characteristics, moving from the top, down to a
more detailed level:

• The IE allows the afloat forces to gain access to national intelligence,

operational intelligence (OPINTEL), and tactical intelligence (TACIN-
TEL) on a worldwide scale.
78 Naval ISR Fusion Principles, Operations, and Technologies

• Ashore nodes include networked MOCs, big data centers, major analyt-
ic centers, and METOC centers. These nodes communicate to deployed
afloat naval units via satellite communications. The nodes interface with
the U.S. Intelligence Community collection sources by the ashore fa-
cilities. The Primary Ocean Prediction System (POPS) combines world-
wide METOC data to provide weather prediction and environmental
models to deployed fleets.
• Afloat fully capable nodes provide afloat force units with an organic
node with maneuver command, battle management, track manage-
ment and effects analysis capabilities. These nodes maintain computing,
analytics, and network distribution management capabilities. They are
maintained on major force units (carrier or expeditionary strike groups,
Joint Force Maritime Component Command groups (JFMCC)).
• Afloat limited nodes are local computing services that include effects
analysis, track management, and battle management capabilities for
their mission (e.g., CG or DDGs on independent operations or missile
defense operations).
• The afloat node uses a common core, the CANES, to provide ISR ser-
vices with plug-ins (application programming interfaces (APIs)) for
DCGS-N, as well as cyber situational awareness (NCSA), Maritime Tac-
tical Command and Control (MTC2), Navy Integrated Tactical Envi-
ronmental System (NITES2), and Naval Operational Business Logistics
Enterprise (NOBLE). The NOBLE family of systems is composed of
the Naval Operational Supply System (NOSS), Naval Aviation Mainte-
nance System (NAMS), and Naval Operational Maintenance Environ-
ment. These capabilities illustrate the breadth of integration of systems
using common services that are cyber secure and auditable.
• CANES provides common hardware and software services in a SOA
architecture across these mission capabilities; all benefit from common
data strategy, standards, and information security.
• The Distributed Common Ground System-Navy (DCGS-N) is a family
of intelligence systems (FOS) that is interoperable with the other service
and agency DCGS systems and provides multilevel security. It includes
collection management, geospatial and signals intelligence analytics,
display, fusion, and dissemination services. It also provides reach-back
capabilities to MOC support centers.
• Closely associated with DCGS services are battle management systems
dependent on precision target locations derived from multiple sensors.
Battle management encompasses many combat systems, for example:
 Naval Information Fusion Systems 79

Aegis Combat System, Joint Land Attack Cruise Missile Defense Elevat-
ed Netted Sensor System (JLENS) services, air operations battle man-
agement services, IO and electromagnetic warfare (EMW) services that
provide the rapid-response capability to integrate kinetic and nonkinetic
(electronic warfare, directed energy) fires.

This naval cloud-based computing enables the development and deploy-

ment of service-based workflows for ISR tasks using the common core services,
as well as custom services unique to a mission area. Yet the incremental deploy-
ment of such a large enterprise is measured in decades because of the scale and
scope of sophistication. U.S. Navy pioneering concepts were developed in the
late 1990s (NCW, ForceNet), development was initiated in the early 2000s,
and incremental deployment was in the following decades [10].
We now consider the example of an ISR fusion workflow that combines
sensor and source data. The ISR fusion capability seeks to maintain an op-
erational picture of the objects of interest in the maritime operating environ-
ment (from seabed objects to overhead satellites) that addresses these questions
needed by the vessel commanders and the fleet commander. The roll-up of
the information to answer the questions in Table 4.2 is provided by the fusion
process with ISR.
We now drop down to a deeper level of description to illustrate the process
of coordinating services to perform a basic function, using a Unified Modeling
Language (UML) diagram to step through a service sequence. The operation
of basic maritime fusion process for surface ships only is illustrated in a UML
sequence diagram (Figure 4.2) that shows one cycle of a series of fusion services
that update the situation, threat, and tasking to ISR sensors [11]. The set of
services are organized across the top and the lines with arrows that proceed
from service-to-service show the messages and data flow that sequences through
the data fusion process. Note that the arrows are transactions across the digital
network and may be routed through many nodes to arrive at the destination
service.
This process shows the data-driven sequence that evokes services when a
new set of sensor observations (from one or more MultiINT sources). A similar
diagram could be shown for query-driven services that are evoked when an ISR
operator queries the systems to analyze data in the object-situation data store
to answer a commander’s question. Note also how the sequence illustrates the
asynchronous nature of the process, and how the services can be distributed
across compute platforms and even across physical platforms.
The circled numbers in the figure correspond to the paragraphs that fol-
low describing the sequence:
80 Naval ISR Fusion Principles, Operations, and Technologies

Figure 4.2 Sequence diagram of SOA-specific scenario.

1. New source data from ISR sensors (a batch from a single sensor, or
multiple sources) is presented to the data aggregator service. In this
example, the batch contains objects (surface vessels observed by sen-
sors) and extracted from surface search radars on ships and aircraft,
and RF signal intercepts of maritime radars, as well as EO and SAR
ship detections from commercial satellites. The data aggregator sends
the batch to the object refine service. Figure 4.3 illustrates how just
 Naval Information Fusion Systems 81

Figure 4.3 Service-to-service message flow over TCP/IP network.

the first message may be relayed from the ISR aggregator service to
the object refinement service through the standard internet using the
TCP/IP Protocol network to reach the object refine service.
2. The object refine service aligns the object observations to a common
time and space reference. It also requests (GET_Obj) the current ob-
jects in the area of coverage of the new objects from the object-situa-
tion storage.
3. Those objects are retuned (RTN_Obj) from the storage and a correla-
tion measurement is computed to determine if the new objects can
update the current objects in time or location and identify or track
dynamics. This service may also update the track of objects (a mov-
ing ship) using a tracking algorithm to associate a new object with an
existing track, and update the track. If an RF ellipse can be uniquely
associated with an object or track, the ID can also be updated. If satel-
lite observations can be associated with any tracks, identity may also
be updated for the tracks. The updated object data message (Update_
Obj) is then sent to the object-situation store to update its state. Note
that the object-situation storage is the source of the common operat-
ing picture presented to commanders, and it is used by C2 simulations
(explained in Chapter 8) to assess alternative COAs.
4. The completion of object updating from this batch is reported to the
data aggregator service; and it requests and update of the situation as-
sessment, based on the new state.
5. Next, the situation assessment service requests the current situation
(GET_Sit) and, when retrieved, estimates the change in situation
based on the new objects, updated tracks, and their relationships.
(This process is described in greater detail in Chapter 6.) The updated
situation is sent to the object-situation store (RTN_Sit).
82 Naval ISR Fusion Principles, Operations, and Technologies

6. Once the situation is updated and reported to the ISR data aggrega-
tor, a REQ_Thrt is issued to the threat assess service to determine if
changes to the situation pose a threat (or offensive opportunity) to
the fleet or its mission. After issuing a GET_Sit and receiving the
newly updated situation (RT_Sit) the threat assess service assesses
the implications of the maritime situation and projected outcomes
from alternative COAs. By requesting a maritime surface simulation
(REQ_SIM) from the simulate service.
7. The simulation service explores the space of current and possible fu-
ture situations over a defined time horizon (e.g., 12 hours) and returns
results (RTN_SIM) to the threat assess service.
8. The threat assess service evaluates (a self-request loop) the potential
threats and issues an Update_Thrt report message to the object-situa-
tion store, as well as the ISR aggregator service.
9. The aggregator service finally requests an update to sensor tasking
(REQ_Tasking) based on the new situation and identified threats that
may change priorities in the ISR network. The resource management
service requests the current situation/threat profile (GET_S/T) and
then computes an updated ISR strategy, sensor allocation, and collec-
tion schedule (tasking) based on the updated situation. The updated
tasking (Update_Tasking) is sent to the object-situation store and to
the ISR aggregator, which forwards it to the appropriate ISR sensor
systems.

Note that this simple example sequence of services only maintained cus-
tody of surface ship threats; a complete seabed-to-space process would maintain
an object-situation store of the complete domain of objects, the full domain
situation, and the threat imposed by the aggregate behavior of all objects.

4.3 Naval C4ISR Challenges

While the preceding sections have outlined the enterprise and information level
concepts for implementing a large naval ISR capability, there remain many chal-
lenges to achieve the scope and scale of such a large enterprise. This is because,
of course, it is not just ISR, but all application systems and the infrastructure
itself that must be translated from stovepipes to the distributed computing en-
vironment. We summarize the challenges in the areas of computational imple-
mentation and operational transformation.
Computational implementation challenges are as follows:
 Naval Information Fusion Systems 83

• Investment and planning for transition. Of course, any navy with a large
investment in legacy ISR stovepipes must develop a transition strategy
with incremental delivery milestones, migration paths, and budgets.
These must guide the introduction of new IT infrastructure, hardware,
software, and translation of legacy capabilities into services (breaking
apart, wrapping, expanding to a shared service useful across ISR mission
threads, etc.).
• Governance policies and standards for data and services. These must be es-
tablished, configuration managed, and adhered to by all components. In
addition, methods must be developed to continually audit SOA systems
to assure compliance.
• Services sequence profiles and metadata management. Methods must be
established to design, document, manage, and test the large number of
messages exchanged between service for all ISR workflows. Profiles must
be developed to enforce compatibility with standards.
• Bandwidth management. The asynchronous and message-passing char-
acteristic of SOA also requires monitoring and management of the ca-
pacity-limited channels in the network.
• Security. Information assurance and security standards must be estab-
lished to support all services; application-level security alone is not ap-
propriate in the SOA environment which is reliant on message passing
across applications using many layers of the computing stack.

Operational transformation challenges are as follows:

• ISR discipline organization. The movement from stovepiped ISR disci-

plines to all-domain introduced the significant challenge to perform dis-
tributed all-domain analysis ISR rather than independent analysis of air
warfare, ASW, surface warfare, electronic warfare maneuver, and more
as independent disciplines. This will also increase the complexity of in-
formation provided to afloat operators. While the single domain analy-
sis disciplines will remain, all source analysts will address coordinated
adversary actions across all domains. For example, preparations for a
coordinated precision air-missile and cyber strike must be recognized by
a warning analyst monitoring all domains. The ISR disciplines (e.g., col-
lection managers, sensor operators, data analysts, intelligence analysts,
operators, and targeting analysts) will address the broader capabilities
of ISR and be able to use tools to that integrate information across all
INTs.
84 Naval ISR Fusion Principles, Operations, and Technologies

• All-domain concept of operations (CONOPS). The concept of operations

for performing analysis and targeting across all domains will also require
new concepts to enable cross-domain coordination, tight integration of
collection and analysis, and rapid delivery of targetable intelligence to
C2. The operational concept of optimal maneuver requires an under-
standing of the high-dimensional operating space (space, air, surface,
undersea, cyber, etc.) and the CONOP must enable all to comprehend
this; it is much larger than the physical common operating picture of
the past [12].
• Training. Of course, with discipline and CONOP challenges, training
will require tailoring to enable the disciplines to teach broader knowl-
edge, develop deeper skills in many complex areas, and apply more criti-
cal thinking abilities to address the large domain of problems they will
face.

The following chapters move toward the implementation of the ISR pro-
cess providing more detail on the required data fusion functions.

Endnotes
[1] NATO is developing ISR technologies similar to those described in this chapter. See
NATO Science and Technology Organization 2021 Highlights, “Interoperability
and Networking of Disparate Sensors and Platforms for Tactical ISR Applications,”
pp. 39–40, and “Data Environmental Knowledge and Operational Effectiveness
(D-EKOE)” pp. 41–42, https://www.nato.int/nato_static_fl2014/assets/pdf/2022/4/
pdf/2021-NATO-STO-Highlights-web.pdf.
The Royal Australian Navy (RAN) is studying robotics, autonomous systems, and AI
(RAS-AI) technologies directly applicable to naval ISR applications. See Slapakova, L., P.
Fusaro, J. Black, and P. Dortmans, “Supporting the Royal Australian Navy’s Campaign
Plan for Robotics and Autonomous Systems: Emerging Missions and Technology Trends,”
Santa Monica, CA: RAND Corporation, 2022, https://www.rand.org/pubs/research_re-
ports/RRA1377-1.html.
[2] Here, we only introduce the SOA concept as an enabler for ISR fusion; for more depth,
see Erl, T., Service-Oriented Architecture: Analysis and Design for Services and Microservices,
Pearson, Dec. 2016.
[3] Joint Publication 1-02.
[4] Management of the Department of Defense Information Enterprise (DoD IE), Depart-
ment of Defense Directive 8000.01, March 17, 2016.
[5] By core fusion services, we refer to the most elemental functions that make up each of the
fusion levels. For example, level 1 fusion refines individual objects from raw data (e.g.,
targets in imagery). The core functions of level 1 for a single source include object detec-
 Naval Information Fusion Systems 85

tion in raw data, object extraction from raw data, alignment of the extracted object in
space and time to a common reference, correlation of the object with other objects and
association of the object measurement with other object measurements, combination of
data from associated measurements, dynamic tracking of object kinematics, and more.
[6] “Web Services Description Language (WSDL) Version 2.0 Part 1: Core Language,” W3C
Recommendation June 26, 2007, http://www.w3.org/TR/wsdl20.
[7] Joint intelligence preparation is a continuous process. See JP 3-0, Joint Operations, 17
January 2017, Incorporating Change 1; October 22, 2018, p. xiv. See also Cross-Domain
Synergy in Joint Operations Planners Guide, Joint Staff Force Development J7; January
14, 2016; https://www.jcs.mil/Portals/36/Documents/Doctrine/concepts/cross_domain_
planning_guide.pdf?ver=2017-12-28-161956-230.
[8] See “Challenges to Security in Space,” Defense Intelligence Agency, January 2019, https://
media.defense.gov/2019/Feb/11/2002088710/-1/-1/1/SPACE-SECURITY-CHAL-
LENGES.PDF.
[9] Tim Anderson, Battlespace Awareness, and Information Operations Program Office
(PMW-12) for Intelligence Analytics 2018 Meeting (Brief ), PEO C4I, Approved for Pub-
lic Release unlimited, February 7, 2018.
[10] The historical background to NCW and ForceNet is provided in the introduction to
Chapter 7.
[11] The Unified Modeling Language (UML) specifies diagrams that represent SOA service
sequences and messages to aid in design and analysis. The sequence diagram is a model
(abstract representation) that is used in model-based engineering. The example is a very
simple, but illustrative, description of the basic data fusion process. See Unified Modeling
Language (UML) Specification Version 2.5.1, OMG, December 2017.
[12] For an Air Force perspective of this challenge, see Lingel, S., et.al., “Joint All-Domain
Command and Control for Modern Warfare: An Analytic Framework for Identifying and
Developing Artificial Intelligence Applications,” RAND Report RR-4408/1-AF, 2020.
5
All-Domain Fusion and Operation
Challenges
The introduction of the term all-domain has created the vision of integration
of technical and operational activities across several domains—spheres of ac-
tivity or influence with common and distinct characteristics in which a force
can conduct joint functions [1]. The term multidomain also exists to describe
systems that operate across multiple domains, but all-domain emphasizes full
integration and synchronization across the entire spectrum of conflict. At the
U.S. DoD level the broadest application is in a joint all-domain operations con-
cept that is comprised of air, land, maritime, cyberspace, and space domains,
plus the electromagnetic spectrum [2]. Actions by the joint force in multiple
domains integrated in planning and synchronized in execution, at speed and
scale needed to gain advantage and accomplish the mission [3]. Ultimately,
all-domain seeks the coordination and optimization of everything, everywhere.
In this book, our focus is on naval maritime and littoral operations; the
emphasis is on integrating and orchestrating the ISRT sensor and C2 opera-
tions for naval and marine forces. While these forces must coordinate with air,
space, and land forces as well as intelligence services, the focus here remains on
naval operations [4]. Coordination of sensing (ISRT), command and action
(C2) across all of the domains of naval operations is envisioned to enable sur-
prise, the simultaneous and sequential application of force at enemy centers of
gravity, and the ability to gain both physical and psychological control over the
operational environment. The key domains we focus on in this chapter include
maritime, air, space, and cyber.
The all-domain concept requires a significant degree of technical and op-
erational requirements to achieve the desired superiority in information across

87
88 Naval ISR Fusion Principles, Operations, and Technologies

the domains, the effectiveness of decision making across those domains and
the impact of coordinated force application. The driving requirements include:

• Operational requirements: (1) Coordinated planning and understanding

of the expected effects of synchronized C2 and force application (e.g.,
physical and psychological effects), (2) distribution of command intent
to allow decentralized and adaptive execution of forces, and (3) speed of
decision, execution, and assessment to conduct coordinated long-range
engagements by the distributed platforms.
• Technical requirements: (1) A global scale information network (also
called a grid or mesh) then enables the rapid and secure exchange of sen-
sor data, sensor control, and C2 data among participant across domains,
(2) data fusion capabilities to correlate, associate, and combine informa-
tion across domains to sustain a timely and accurate operating picture
across domain, and (3) decision support to aid commanders to under-
stand, maneuver, and apply combat power appropriate for the military
objectives.

The challenges all relate to the ability to carry out the maritime ISRT
sequence that moves from surveillance-to-targeting that follows a typical se-
quence of states [5]:

Detection: The wide area surveillance process performs area scanning and
directed area searches (often based on cues from other sources such as a
port-departure notice) with the objective of detecting targets. For surface
search, vessels are detected by zone of operation, and are categorized as
cooperative (reporting ID by declared ID on the AIS per international
law), or dark (detected but not emitting or identified—unknown vessel).
For air search, aircraft targets on radar or targets emitting radar or other
signals detected by ELINT are detected.
Identification: Once detected, surface vessel must be classified by vessel
type and uniquely identified (e.g., by hull, name) to declutter the surface
picture. If cooperative AIS data is not associated with a vessel, other non-
cooperative methods such as vessel RF signal fingerprints (specific emit-
ter identification (SEI)) must be used. For aircraft, efforts are made to
identify the target by commercial track and transponder reporting, iden-
tification friend or foe (IFF) interrogation, flight behavior, flight path,
point-of-departure (airfield), or other noncooperative means. Identifica-
tion is critical to understand both the surface and air picture, and the po-
tential implications of an adversary’s combined air-surface maneuvering.
The ISR system must assess the risk of target behaviors, including the
 All-Domain Fusion and Operation Challenges 89

vessel history and particulars (ownership, cargo) for identified vessels and
unidentified vessels. This state involves performing follow-up tasking to
sensor systems to resolve identification and initiate tracking.
Track: The next state of surface or subsurface vessel ISR is to maintain a
track throughout passage in an area of interest—also known as maintain-
ing custody of the vessel. This requires the sustained observation of the
target by continuous or periodic sampling to assure the target is not lost
in clutter. An emitting AIS target, easily tracked, may go dark, causing
the ISR system to lose track if it cannot switch to another sensor mode
to regain custody. Targets in-track are correlated across sensors and inte-
grated to provide a common operating picture (COP). Air targets likewise
require a sensor (e.g., radar) to sustain periodic observations (e.g., radar
returns) of the target at a higher rate than search to maintain a track of
the aircraft trajectory. The track can be represented by a series of observed
POSITs or a computational tracking model (e.g., a Kalman filter) that
smooths the POSITs, estimates the trajectory parameters, and predicts the
expected location of the next sample observation. Throughout tracking,
the ISR system monitors track behavior for anomalies to alert to unusual
or threat-like behaviors, or maneuvers to risk loss of track.
Targeting and engagement: The nomination of a target for targeting refers
to the selection of a target for action. This action may be to intercept
a vessel for reconnaissance or boarding-inspection, or for committing a
weapon to engage an air or surface target.

The following four sections (Sections 5.1–5.4) describe the challenges

that these requirements impose. The challenges fall into four areas. The first
area includes the effects of the distribution of sensors, C2, and weapons across
spatial distances, and the scale of the network. Travel time for information
across these distances introduces latency (time delay) that must be addressed by
any networked system. But latency is also introduced in traffic as network scale
increases because network congestion and interference also contribute delays.
Latent data arrival to data fusion nodes has the potential to introduce signifi-
cant data association errors. (This topic is addressed in greater detail in Chapter
7; distributed data fusion systems must be able to adapt to the inevitable laten-
cies in networked systems.)
The second area of challenge is the ability to maintain an understanding
of the dynamic situation when using many sensors. The sensor is effectively
sampled when performing surveillance and tracking of maritime targets and
the required update rates are inherently related to the rate of change of the dy-
namic targets. While the first challenge areas are based on spatial and temporal
requirements, the third area is driven by the measurement accuracy demands
90 Naval ISR Fusion Principles, Operations, and Technologies

of fire control systems to deliver physical and cyber weapons. The fourth area
addresses the need for integration of cyber capabilities to assure commanders
understand adversary coordination of attacks in this domain with the physical
domain.
These challenges in the following sections inherently call for speed and
precision, and in turn, may necessitate degrees of automation and augmented
cognition to achieve high levels of performance. The potential contributions
of automaton and augmentation are further described in Chapters 6, 7, and 8.

5.1 Challenge 1: Spatial Distribution, Association, and Latency

The distribution of maritime operations distributes naval platforms and there-
fore the sensors, weapons, and communication relays are spatially distributed at
wider distances than traditional battle groups formations or carrier groups that
operate in clusters to concentrate offensive fire power, defensive measures, and
overlapping sensors.
This spatial distribution contributes following operational challenges to
ISRT:

• Organic sensing gaps: The wide distribution of surface platforms intro-

duces gaps between platforms eliminating organic sensor overlaps (sur-
face and air radar coverage, towed array acoustic coverage). These large
gaps then must be covered by nonorganic sensors to prevent vulnerabili-
ties to air, surface, and submarine threats. Consider the simple example
(Figure 5.1) that illustrates the relative density of fleet vessels in three
cases: (a) a carrier battle group tightly spaced for concentrated force, (b)
the same force distributed over a moderate distance, and (c) same force
widely distributed with very large gaps in sensor coverage. The demand
for ISR coverage to monitor these gaps can be significant—especially in
seas where heavy commercial traffic creates significant clutter to chal-
lenge the ability to maintain custody of adversary ships and submarines
that enter the operating area. It is reasonable to consider, in the wide
case C, that the fleet is distributed over a vast 250 nm (radius) circular
area of operations with an area coverage of 200,000 nmi2 [6]. In the
figure, we consider a DMO fleet (uniformly distributed) consisting of 1
carrier strike group (1 carrier, 8 destroyers, 2 maritime patrol aircraft, 2
air surveillance aircraft, 5 long-range UAVs, 10 deployed Seabed Sensor
networks, 10 UUVs, 10 USVs, and 2 attack submarines). There are 31
vessels in this fleet and 9 available air assets; 5 in the air at any time; this
results in 36 fleet sensing platforms. Figure 5.1 also considers larger fleet
sizes to show the intervessel distances for 72 and 108 platform fleets.
 All-Domain Fusion and Operation Challenges 91

Figure 5.1 Vessel density at three levels of distribution.

• Distributed sensing: DMO necessitates a greater reliance on distributed

and cooperative sensing, where sensors must communicate and share
their reports while enabling control of their operation from distant us-
ers. The concept of organic sensors (sensors controlled by the host plat-
form) must give way to shared and orchestrated sensing to provide a
common operational picture for all.
• Domain distribution: Of course, the breadth of operations from the sea-
bed to space introduces additional dimensions of distribution beyond
spatial distribution. Attacks will be coordinated across domains to create
a maximum disruptive effect, or a highest quality (believable, seductive)
deception. For example, cyber and seabed sensors must be coordinated
to distinguish anomalies in reported sensor data and disruption of net-
work channels that may indicate a coordinated attack on a critical un-
dersea sensor network and its associated communication channel.
• Target dispersal and weapon range: DMO also causes opposing naval
forces to distribute more widely to confront the distributed fleet, and
this in turn demands greater ISR coverage to monitor the threat and
longer-range weapons (missiles, UAV loitering, etc.). The DMO force
is driven to apply longer-range missiles, for example, to engage targets
well beyond the horizon, and beyond organic sensor ranges. These long-
range missiles (Figure 5.2) must rely on cooperating nonorganic sensors
to enable fire control for the missile in flight. We discuss this further in
Section 5.3.
92 Naval ISR Fusion Principles, Operations, and Technologies

Figure 5.2 Long-range ASM trajectories.

These operational factors, due to the spatial distribution of targets and

sensors, introduce technical challenges that we address throughout this book.
The distribution of naval vessels (and the adversary distributed response) intro-
duces two key challenges that must be addressed by technical solutions:

1. Latency: The networking of sensors to relay nodes and to fusion nodes,

and then to weapon nodes, introduces time delays—latency between
the time of detection of a target and time reported. As distribution
increases, and network scale increases, even latency increases. Latency
does not increase linearly with distance, but nonlinearly with the paths
and nodes across a network. For example, the path from an autono-
mous surface sensor to a communication relay network of LEO satel-
lites and then to a local destroyer may vary depending on the location
of satellite relays. The solution to this challenge is a dynamic network
that maintains shortest-path communication from sensors to fusion
nodes, by adapting as sensors and fusion platforms move; of course,
high-speed routing of the data in the network is necessary.
2. Report association: The latency between detection and reporting, and
the distribution of sensors reporting at different observation (or sam-
 All-Domain Fusion and Operation Challenges 93

pling) times poses a challenge to properly associating reports—assur-

ing that reports are properly assigned to common targets [7]. The so-
lution to association is based on assuring appropriate sampling rates
and applying context and target signature data (e.g., shapes, signals,
and behaviors) to assure the correct matching of data across sensors
and from observation to observation. We discuss the technical ap-
proaches to perform association in Chapter 7.

5.2 Challenge 2: Temporal Sample Rate and Dynamic Targets

Naval vessels (ships, submarines) and naval aircraft maneuver—they are in con-
tinual readiness to maneuver dynamically as an element of fleet tactics in war.
Captain Wayne Hughes emphatically points out:

[Naval] warfare is deadly conflict. Tactics, as the devices by which battles

are waged, are conceived, and executed at the center of this violence…
They are more visceral in the execution than policy, strategy, operational
art, or logistics…tactical plans and battlefield decisions are influenced by
an environment of controlled violence and directed chaos… At sea the
predominance of attrition over maneuver is a [basic principle]… Forces at
sea are not broken by encirclement, they are broken by attrition [8].

Hughes points out that naval warfare is so destructive and decisive that
the avoidance of it altogether weighs heavily on strategists. For this reason, fleets
are dynamic and maneuvering at peace and war to avoid detection and tracking
and to prepare for immediate action.
Therefore, a pillar of naval strategy is the necessity that ISR capabilities
maintain continual custody of the location, movement, and activities of op-
posing naval forces (in-port, maneuvering at-sea, hiding on-station, and in-
maintenance). For the naval fleet conducting DMO operations this demands a
persistent tracking of adversary fleets and knowledge of the commercial mari-
time clutter in which they operate.
By tracking, we refer to the ability to maintain a record of the successive
positions of a moving object. We distinguish the course that may be planned by
the object (e.g., a commercial shipping route) and the track that is the actual
path or trajectory followed by the object. It is helpful to further distinguish two
primary methods of performing the tracking process (Figure 5.3):

1. A time series of POSITS records the track of a ship, for example. The
common maritime track produced by a series of AIS POSITS can plot,
in near real-time, the trajectory of a ship. AIS Class A transponders
typically provide a new location update every 2 to 10 seconds while
94 Naval ISR Fusion Principles, Operations, and Technologies

Figure 5.3 Approaches to physical target tracking.

underway and every 3 minutes while at anchor, enabling a persistent

track by connecting the POSITS with a mathematically smoothed
curve.
2. Computational track models are developed for ships and aircraft by
filtering the observations from radars and other time sampled sen-
sors. The filtering mechanism (also called a tracking loop) develops
over time a computational model of the behavior of the target under
track. In Figure 5.3, a typical filter mechanism is shown where: (1)
observations are received and compared to the model prediction for
that observation, (2) the comparison provides the error in the model
estimate, (3) an update of the estimated current state, and (4) a predic-
tion of the next expected observation state. The computational model
adapts as it estimates the current state, predicted next state at the time
of next observation, and the model error (covariances between state
variables) to provide a measure of the track quality. Several sophisti-
cated estimator mechanisms have been developed and performance is
heavily dependent on the sensor sample rate, the target dynamics (ma-
neuverability), the density of targets, and clutter. Most estimators are
statistical estimators that rely on a knowledge of the statistical proper-
ties of target dynamics [9].

Traditional maritime tracking is performed by organic sensors located on

a surface ship looking out to the horizon to track at sample rates between 0.5
 All-Domain Fusion and Operation Challenges 95

samples/sec and 1 samples/sec (surveillance) or >10 samples/second (racking a

maneuvering air target).
The challenge for DMO is to perform tracking of targets at great distanc-
es, by distant platforms applying long-range networked sensors. These sensors
may be on unmanned platforms (e.g., satellites, UAVs, and USVs). The degree
of difficulty to perform tracking in these cases and the approach to sensing is a
function of the target speed and the sample rate of observations.
Consider the wide range of target speeds for important naval targets to
track (Table 5.1); the table does not consider the maneuverability of each class
of targets that must also be considered as turning and maneuvering poses an
additional challenge to sustained tracking (and adversaries know this).
It is important to recognize that the table is for physical kinematic targets
that are important to the fleet; but nonphysical targets in the cyber and informa-
tion domains are also critical. In these domains the fleet must also be able to
track the flow of cyber-attacks and the propagation of information threats that
spread through the cyber medium to influence human targets, decision making,
and actions. We discuss this further in Section 5.4.
There are many factors that influence the desired ISR sample rate, includ-
ing the purpose of tracking, the interval between observations (the inverse of
the sensor sample rate), target clutter density, and target maneuvering rate. We
can illustrate the challenges by considering the guided missile frigate (FFG)
ship track in Figure 5.4. Four different tracking challenges are posed to tracking
as the vessel departs a naval base and heads to blue water to intercept a target.
The FFG departs port (1) proceeds to a loiter area to perform a loop timing
maneuver and (2) to rendezvous with an escort submarine before heading to

Table 5.1
Relative Speeds of Targets to be Tracked
Desired Sample
Max Velocity Rate to Track
Target Max Velocity (ft/sec) (samples/min)
Container, commercial ship 20–25 kts 33–42 fps 1
CVN (nuclear aircraft carrier) 30 kts 50 fps 2
DDG (guided missile destroyer) 35 kts 60 fps 3
Submarine (submerged) 30–40 kts 50–68 fps 3
Fast missile corvette 30–60 kts 50–116 fps 5
Aircraft (patrol) 350 kts 600 fps 20
Cruise missile (subsonic) 550–700 mph 800–1,000 fps 30
Aircraft (fighter) 1,500 mph 2,200 fps 60
Hypersonic missile (Mach 7) 3,800 mph 5,575 fps 60–120
ASM cruise missiles such as Excoct or Harpoon.
Hypersonic boost-glide vehicle and cruise missiles range from 5 to 7 Mach.
96 Naval ISR Fusion Principles, Operations, and Technologies

Figure 5.4 Tracking an FFG on an operational intercept mission.

the operational area. Once the FFG and submarine are ready, the FFG departs,
crossing the offshore shipping lanes and performing maneuvers close to mer-
chant traffic behaviors tracks to confuse efforts to track the vessel (3). Once
outside of the shipping lanes, the FFG proceeds at 30 kts to the operational area
to prepare for intercept and engagement with an opposing force. In the process
of that operation, the FFG and other combatant vessels joining the operation
make several tight and crossing maneuvers (4) to evade tracking.
Tracking the FFG in this scenario requires, first the detection of the ship
departure (1) from the port (by AIS signal, seabed or surface sensors, or other
ship, air, or space radar sensors) and initiation of a track process that requests
persistent sequential sensing to sample the track at a sufficient rate to maintain
initial custody. As the FFG loiters for rendezvous (2) the tracking may require
increased sampling to keep up with the turns. As the FFG enters the ship lanes
(3) and maneuvers back and forth to blend with merchant traffic to cause any
tracker to follow a merchant ship and lose the combatant FFG. (This is also
a period that the AIS may be turned off to cause the FFG to go dark, to AIS
tracking.) Finally, after proceeding to the operational intercept point, the FFG
may again perform sharp maneuvers (4) to cause a tracking process to break
lock and lose custody.
AIS reporting rates at 6–30 samples per minute provide excellent sampling
and a high-altitude UAV may provide a similar sampling rate on this target, but
OTH radars and satellite sensors may provide much lower sampling rates. Dis-
tributed sensor systems that integrate observations from multiple sensors, con-
tributing at different times, may provide uneven sampling further complicating
track custody. The relation between target separation and sampling interval on
tracking performance for the FFG track in Figure 5.4 is illustrated in Figure 5.5
[10]. The ability to correlate samples is unambiguous when the ship is widely
separated from ship traffic and is broadcasting a high-rate AIS signal. As the
ship conducts loitering maneuvers in low density areas (Region 1) the perfor-
mance diminishes if the sample rate cannot distinguish the maneuvering rate,
and in high maneuvering at lower sample rates (due to AIS being turned off and
observation rates from air and space sensors are reduced). In Region 4 the track
 All-Domain Fusion and Operation Challenges 97

Figure 5.5 Regions of correlation and tracking performance.

may be lost, unstable, or misclassified. Throughout the passage through the

shipping lane (Region 3), the background clutter of merchant ships reduces the
separation between targets and can cause incorrect assignment of observations
to tracks and, ultimately, loss of track.
In summary, the factors that influence tracking performance include:

• The purpose of tracking: If the purpose is loose surveillance and custody

over a long period, lower sample rates are acceptable, but higher sample
rates are required for evading and maneuvering targets, and for targeting
and engagement.
• The interval between observations (the inverse of the sensor sample rate)
is important to maintain track continuity and is a function of the target
rate and rate of maneuver.
• Target clutter density is the number of nontarget ships in the sensor field
of view; high densities challenge the tracker to distinguish between ob-
servations when tracks cross or closely cotravel.
• Target maneuvering rate can cause the tracker to lose track when an
observation does not meet the predicted next observation criteria; this
requires the tracker to rapidly adapt the behavior model to determine
the target has abruptly changed behavior (e.g., cruise behavior to rapid
maneuverings).
98 Naval ISR Fusion Principles, Operations, and Technologies

These operational factors introduce three key requirements that must be

addressed by technical solutions:

1. Sensors must provide coverage to enable sample rates sufficient to

maintain tracks on combatant targets. This means that they must
have sufficient field of regard and pointing agility to assure sampling
across the anticipated density and dispersion of maritime targets. The
technical solution requires a network of distributed sensors that can be
coordinated to sample multiple targets in time sequence to maintain
multiple tracks, adapting the sample rate as the targets transition from
cruise to maneuver.
2. A robust network between sensors assures that sensors can be orches-
trated to control pointing to targets, perform cueing and handoff from
sensor to sensor, and provide paths for sensor data to appropriate fu-
sion noes where data are associated.
3. Computation to perform sensor management for the functions of cue-
ing, search, handoff, and tracking based on track models that guide
sensors to the predicted next observation of a track.

5.3 Challenge 3: Accuracy for Fire Control and Missile Engagement

Beyond tracking, the most stringent demand for accuracy of the state of a sur-
face target and its predicted states are for fire control to enable DMO missile
engagements. In this section we illustrate fire control configurations to conduct
long-range surface to surface missile (SSM) engagements by networked sensor
and command links, targeting by remote control from OTH.
Traditional targeting has been performed by sensors that are organic to a
surface ship. The ship’s surveillance radar may acquire an air or surface target
by line-of-sight contact and then dedicate a fire control radar to acquire and
track the target to provide a high accuracy, high sample rate track of the target.
Using this track the missile is launched and the ship provides required inflight
guidance to the missile through flight until the missile acquires the target with
its seeker sensors.
For OTH targeting in DMO, networked sensors must provide the sens-
ing with sufficient accuracy to engage a distant target. The term integrated fire
control (IFC) refers to the means to integrate fire control mechanisms (sen-
sors, C2, engagement control, guidance) and CEC refers to the mechanisms
that enable sensor platforms, firing platforms, and C2 platforms to conduct
coordinated missile engagements (air and surface targets). In these cases, the
demand for accuracy of the target track up to and after launch of the missile is
paramount. The accuracy requirement:
 All-Domain Fusion and Operation Challenges 99

• Provides accurate coordinates of the target at time of launch and pre-

dicted location at time of missile arrival to optimize the trajectory to the
target area;
• Provides accurate inflight updates as the target maneuvers when it re-
ceives an inbound missile warning; inflight updates provide the predict-
ed location at missile arrival.
.

In this section three conceptual CEC modes appropriate for long-range

surface engagement in DMO are illustrated. Figure 5.6 and the following para-
graphs describe three operations of these modes using remote OTH sensor data
from satellite, UAV, and USV sensor systems [11]. More modes of operation
have been conceived including: Precision Cue (where a remote sensor provides
a tip to a potential threat and cues other sensors to acquire a track and pre-
pare to engage if appropriate; in Figure 5.6, this may precede all of the modes
shown), Launch on Remote (where the launch is entirely based on remote data
from single remote sensor or the composite of multiple sensors), and Preferred
Shooter Designation (where firing units are selected from a group to optimize
the engagement). While earlier concepts published were focused on air target
CEC, here, we describe CEC modes suitable for DMO SSM engagements [12].
Mode 1: Engage on Remote. In this mode one or more remote sensors pro-
vide the cue to the remote target, as well as the identification and tracking data
to enable commander decisions to engage the target. The sensors also provide
the track update data to commit to the engagement and deliver inflight updates
that are relayed to the missile throughout the engagement data upon which all
(or portions) of an engagement is conducted. The relay network provides the
sensor-to-fusion node data and relays the inflight data to the weapon. The fu-
sion node combines multiple sensor measurements (also called composite data)
into the common track and ID used for firing decisions and engagement. This
mode requires the composite track to maintain fire control quality information
(accuracy and accurate target maneuver projections) throughout. Notice that
the remote sensors provide all data directly to the firing vessel.
Mode 2: Forward Pass. In this mode, a C2 vessel controls the engage-
ment, and remote platforms closer to the target (the middle panel in Figure
5.6, a UAV and USV) provide in-flight updates to the weapon. The missile is
launched from a firing vessel and is handed off (forward passed) to the UAV
and/or USV platforms to provide terminal guidance data. The remote sensor
network (space plus UAV-USV platforms) provides fire control quality target-
ing data to firing unit. The firing unit launches the weapon by command from
control unit and missile updates are provided by the close-in (and covert) un-
manned platforms through the relay network.
100 Naval ISR Fusion Principles, Operations, and Technologies

Figure 5.6 Conceptual CEC modes.

Mode 3: Remote Fire. In this mode, the firing unit is commanded to

fire with direction from the C2 vessel and is not involved in fusing data or
 All-Domain Fusion and Operation Challenges 101

maintaining track of the remote target; it is provided the launch and target data.
The remote sensor network detects the target, and the command vessel receives
data and completely manages the fusion of data, sensor tasking, and engage-
ment by the remote firing unit. Missile updates are provided by the remote
sensors (e.g., UAV or USV) as directed by the control unit.
As missile speeds move from subsonic cruise missiles to supersonic cruise
and even hypersonic, the much shorter missile time-of-flight allows much less
time for surface targets to maneuver (Table 5.2) and reduces the predicted track
location at the time of missile arrival. For example, a ship traveling at 30 kts can
traverse almost 10 miles in the time it takes a 700-mph cruise missile to traverse
a 200-mi range-to-target. On arrival, the missile sensor (seeker) searches for the
target and verifies signature attributes before guiding to impact. The accuracy
of the inflight track updates is critical to enable the missile seeker to acquire the
target in the endgame—the final 10s of seconds of flight.
Demonstrations of long-range CEC surface-surface engagements in early
2021 by the U.S. and Russian navies were conducted to attack targets over 300
km by ship-launched missiles using cooperative engagement strategies [13].
The U.S. experiment used a blend of remote sensor data from manned and
unmanned ships and aircraft to provide fire control information to a guided
missile destroyer, enabling the launch of an antisurface missile from OTH to
hit a target more than 250 miles distant [14]. Rear Admiral James Aiken (U.S.
Navy) said of the demonstration, “We teamed manned and unmanned vessels
together. We also used the fusing capability… It was totally passive where we
didn’t have active sensors on target …We also look for space as well to identify
the target and then once we found the target, we were able to track it because
of the [electromagnetic signal] that was coming off the target, develop lines of
bearing, then launched the missile” [14].
The implementation of these concepts requires the consideration of two
additional complexities for operational applications. First is the challenge of

Table 5.2
SSM Time-of-Flight at Long-Range Engagements
Missile Time-of-Flight to Target (Minutes)
Engagement Distance (nm)
Missile
Missile Type Speed (Mph) 100 200 400 600
Cruise 500 12.0 24.0 48.0 72.0
Cruise 600 10.0 20.0 40.0 60.0
Cruise 700 8.6 17.1 34.3 51.4
Supersonic 1,000 6.0 12.0 24.0 36.0
Hypersonic Mach 5 3,000 2.0 4.0 8.0 12.0
Hypersonic Mach 7 5,000 1.2 2.4 4.8 7.2
102 Naval ISR Fusion Principles, Operations, and Technologies

hypersonic weapons that pose severe requirements for tracking that are required
by the defender. Sensors must track the maneuvering boost-glide weapons
through a plasma sheath and provide sustained high accuracy tracking to enable
intercept. The time-to-intercept is very short and national efforts are required
to develop intercept capabilities.
The second complexity is the operational need for missile salvo analysis to
determine the appropriate number of fires (a salvo) necessary to achieve target
destruction. Typical operational analyses must consider at least the following
factors to determine salvo sizes [15]:

• Missile single-shot kill probability (SSPK or PKSS);

• Effect of target deceptive measures (signature deception, deceptive ma-
neuvering, decoys, etc.);
• Effect of close-in weapon systems designed to disable the missile by cy-
ber, electromagnetic, or cyber means.

These operational factors introduce two key challenges that must be ad-
dressed by technical solutions:

1. Prioritize and orchestrate the net at critical times, particularly during

critical tracking periods prior to and during engagement. The system
must be able to project ahead and assure that sensors and network
access will be available (especially space and unmanned platform sen-
sors) during critical tracking and transition to engagement and missile
flight stages.
2. Position and orchestrate distant sensors to achieve critical accuracy
and ID during these periods. Unmanned sensor platforms must be
positioned, and sensors pointed to provide the necessary data to sup-
port engagement.

The solution to these requirements requires productive models that fast-

simulate the engagement and orchestrate sensing and relay channels to assure
engagement continuity of service.
In conclusion, it is important to recognize that humanitarian organiza-
tions have examined the legal issues raised by remote attacks at long ranges
employing capabilities such as cooperative engagement with advanced weap-
ons. While not technical, the operational considerations of the principles of
targeting law for long-range engagement must be developed (rules of engage-
ment that consider criteria for distinction, discrimination, proportionality, and
precaution) [16].
 All-Domain Fusion and Operation Challenges 103

5.4 Challenge 4: Integrating Cyber Capabilities

While the prior three challenges are driven by the physical properties of time,
space, and kinematic behaviors, the challenge of integrating cyber defense and
cyber offensive capabilities poses a different kind of operational need. Offensive
adversary cyber activities (e.g., probes, disruptions, access, and exfiltration) oc-
cur in seconds to milliseconds (or less) and the decision response must be in a
similar time frame. The cyber and EW targets are nonphysical, seeking to attack
information and ultimately decision making in the minds of humans (analysts,
operators, and commanders).
Of course, a primary target of the DMO fleet will be its networking ca-
pability; as the network degrades, so does distributed lethality, and the widely
distributed vessels become more vulnerable as they must fall back to their or-
ganic sensors.
The DMO fleet will be challenged by two categories of cyber threats:

• Ad hoc and continuous threats are continual external cyber activities that
probe, test, and attempt to disrupt the network through any point of ac-
cess on communication links or computer nodes afloat or ashore.
• Advanced persistent threats (APT) that “…possesses sophisticated levels
of expertise and significant resources which allow it to create opportuni-
ties to achieve its objectives by using multiple attack vectors (e.g., cyber,
physical, and deception). These objectives typically include establishing
and extending footholds within the information technology infrastruc-
ture of the targeted organizations for purposes of exfiltrating informa-
tion, undermining or impeding critical aspects of a mission, program, or
organization; or positioning itself to carry out these objectives in the fu-
ture. The advanced persistent threat: (i) pursues its objectives repeatedly
over an extended period of time, (ii) adapts to defenders’ efforts to resist
it, and (iii) is determined to maintain the level of interaction needed to
execute its objectives” [17]. Will the fleet go to sea with a covert APT
present? In spite of all the security processes in place, it must be prepared
for just such an eventuality.

Consider just a small set of the attack surfaces (categories of attack entry)
and vectors (specific targets of access to the network) that must be protected
by the fleet cyber team (Table 5.3). The threats are diverse (ranging from per-
inserted malware in the supply chain to real-time targeted service denial) and
dynamic as attacks may change throughout an engagement period. The cyber-
attacks can be expected to be coordinated with attacks in other domains to
stress the use of the network for ISR and C2 activities.
104 Naval ISR Fusion Principles, Operations, and Technologies

Table 5.3
Example Naval Cyber Attack Categories
Cyber Attack Attack Vectors Employed
Surfaces by Adversaries Description/Example
Network Network probing and Probe network to map the architecture and
Computing mapping properties of any aspect of the fleet nets
Access leapfrog from open Exploit bridge vulnerability between networks,
internet to secure maintain an advanced persistent threat
mechanism
Denial of service (DoS) by Insert large numbers of targets electronically or
overwhelming input physically (e.g., decoy small vessels, and UAVs)
Malware insertion Insert malware via exploited access channel or
(electronic) persistent
Malware insertion (physical; Insert malware into commercial components
supply chain) (hardware, firmware, software) in the supply
chain
Network Comm link disruption/denial Jam network links (e.g., low signal satellite links)
Communications (jamming)
Links Comm link denial (relay Attack or jam physical relays (e.g., satellite, UAS
destruction) nodes)
Comm link exploitation Analyze network traffic to exploit comm patterns
(traffic analysis)
Network- Network personnel social Social engineer cyber administrator personnel to
Related engineering inadvertently expose information
Personnel Spear phishing attack on Broadcast dedicated spear phishing attacks to
ashore admins seduce low-level net personnel to ingest malware
Error-inducing attacks on net Induce an error in configuration by inserting
personnel controlled instructions to net administration
personnel
False data insertion across Insert data to reduce network or ISR (user)
network personnel confidence in the network and its
application
Physical Attack Confusion-inducing attacks Create critical network conditions (e.g., physical
on net personnel node attack or electrical power disruption),
induce network actions that expose the net to
exploitation

While the fleet conducts cyber defense against such threats, it also con-
ducts cyber ISR to explore the adversary cyber operating environment that
poses threats [18]. The mechanisms of computer network exploitation (CNE)
are employed to passively and actively engage adversary networks to understand
the cyber operating environment. The concept of persistent engagement is where
cyber interactions with adversaries are ongoing, at a level below armed con-
flict, but not simply responsive. It embraces forward defense where a persistent
contest (detect, parry, counter-prob, etc.) is ongoing between cyber adversaries
[19].
 All-Domain Fusion and Operation Challenges 105

Because of the ubiquitous presence of the cyber domain, ashore cyber an-
alysts support this activity from afar as well as afloat analysts that are physically
close-in to threats and may have greater access to aspects of adversary networks.
Ashore-afloat coordination of cyber analysis and attack is critical to achieve an
integrated response to all levels of an adversary organization structure.
But cyber cannot be considered an independent domain—adversaries
may compound cyber-attacks with assaults in other domains. Defensive counter-
cyber operations refer to those activities that prepare for and respond to adversary
attacks across all surfaces. For example, an adversary may coordinate network
denial actions (jamming comm links) with a network computing disruption
(denial of service (DoS)) to limit ISR the ability to observe a critical Red fleet
maneuver. The adversary may further enhance the effect by coordination with
a SIGINT deception that produces distracting decoy signals that will cause ISR
assets to focus away from the critical Red maneuvering. Cyber awareness, in
this case, must be fully coordinated with SIGINT and other sources to detect
the actions, infer the denial/disruption and deception intentions, and respond
with countermeasures.

Endnotes
[1] Joint Publication (JP) 3-0, Joint Operations, describes the operational environment as en-
compassing the physical domains of air, land, maritime, and space, the IE, which includes
the cyberspace domain, and the electromagnetic spectrum. It also describes the joint func-
tions as related capabilities grouped to help commanders integrate, synchronize, and direct
operations. The joint functions are C2, information, intelligence, fires, movement and
maneuver, protection, and sustainment.
[2] The U.S. DoD envisions a Joint All-Domain Command and Control (JADC2) archi-
tecture to enable commanders to (1) rapidly understand the battlespace, (2) direct forces
faster than the enemy, and (3) deliver synchronized combat effects across all domains.
See Hoehn, J. R., Joint All-Domain Command and Control (JADC2), Congressional
Research Service, IF11493, updated July 1, 2021.
[3] Air Force Doctrine Note 1-20, USAF Role in Joint All-Domain Operations.
[4] Each service has a unique role and perspective on its own contribution to all-domain
operations, for example, see the U.S. Air Force perspective in Air Force Doctrine Publica-
tion (AFDP) 3-99, Department of the Air Force Role in Joint All Domain Operations
(JADO), October 8, 2020.
[5] Galdorisi G., and R. Goshorn, “Maritime Domain Awareness: The Key to Maritime Secu-
rity Operational Challenges and Technical Solutions,” in Proc. 11th ICCRTS, September
2006.
[6] For reference, the South China Sea is a marginal sea that is part of the Pacific Ocean,
encompassing an area from the Karimata and Malacca straits to the Strait of Taiwan of
around 3,500,000 km2 (1,400,000 mi2). The China Sea Basin has a maximum depth of
5,016m.
106 Naval ISR Fusion Principles, Operations, and Technologies

[7] In data fusion terminology, we define correlation as the process of measuring quantitatively
the similarity between two measurements (e.g., same lace, same time, same observed at-
tributes) and association as the decision process of assigning two or more measurements to
a comm on target. Correlation precedes association, and association preceded combining
attributes or fusion of data.
[8] Hughes, W., and R. Girrier, Fleet Tactics and Naval Operations, Annapolis, MD: Naval
Institute Press, 2018, pp. 18–19.
[9] See Blackman, S., and R. Popoli, Design and Analysis of Modern Tracking Systems, Nor-
wood, MA: Artech House, 1999.
[10] This figure is adapted from the fundamental regions of tracking performance developed in
the classic textbook on tracking. See Blackman, S. S., Multiple-Target Tracking with Radar
Applications, Dedham, MA: Artech House, 1986, pp. 12–13.
[11] The CEC and integrated fire control capabilities have been evaluated in experiments since
the 1990s. The descriptions are based on the following relevant documents: Grant, C. J.,
“CEC: Sensor Netting with Integrated Fire Control,” Johns Hopkins APL Technical Digest,
Vol. 23, No. 2 and 3, 2002, pp. 149–161. Young, B. W., “Future Integrated Fire Control,”
Proc. 10th International Command and Control Research and Technology Symposium, June
2005. Deering, V., et al., “Open Architecture as an Enabler for FORCEnet,” NPS-
SE-06-002, Naval Postgraduate School, September 2006.
[12] Young, B. W., “Future Integrated Fire Control,” Proc. 10th International Command and
Control Research and Technology Symposium, June 2005.
[13] “Russia says it successfully tested hypersonic missile praised by Putin,” Reuters, July 19,
2021, https://www.reuters.com/world/europe/russia-conducts-ship-based-hypersonic-
missile-test-ifax-cites-defence-ministry-2021-07-19/.
[14] LaGrone, S., “Unmanned Systems, Passive Sensors Help USS John Finn Bullseye Target
with SM-6,” U.S. Naval Institute, April 26, 2021, https://news.usni.org/2021/04/26/
unmanned-systems-passive-sensors-help-uss-john-finn-bullseye-target-with-sm-6.
[15] For a complete discussion of missile salvo and fleet exchange equations. See Hughes, W.,
and R. Girrier, Fleet Tactics and Naval Operations, Annapolis, MD: Naval Institute Press,
2018, Chapter 13.
[16] Boothby, W., “Some Legal Challenges Posed by Remote Attack,” International Review of
the Red Cross, Humanitarian Debate: Law, Policy, Action, New technologies and Warfare,
Vol. 94, No. 886, 2012, pp. 575–595.
[17] NIST. Managing Information Security Risk: Organization, Mission, and Information
System View. SP 800-39 (2011).
[18] Votipka, D., et al., “ISR and Cyberspace,” Air University, 6 July 2017, https://www.
airuniversity.af.edu/CyberCollege/Portal/Article/Article/1238539/isr-and-cyberspace/.
[19] Nakasone, P. M., “A Cyber Force for Persistent Operations,” Joint Forces Quarterly 92,
2019, pp. 10–14. See also Healey, J., and S. Caudill, “Success of Persistent Engagement in
Cyberspace,” Strategic Studies Quarterly, Spring 2020, pp. 9–15.
6
Maritime MultiINT Fusion Processes
In Chapter 4 we enumerated the many state-of-the-art systems applied across
advanced navies to integrate organic and nonorganic sensors and sources, gen-
erally integrated by ISR mission (antisubmarine warfare, surface warfare, antiair
warfare, etc.). In this chapter we introduce concepts and technologies to signifi-
cantly extend the integration and automation to a greater degree beyond what
we described in Chapter 4—including processes to estimate the entire maritime
situation across the missions and the domains of space, air, surface, subsurface,
and the seabed.
The concepts of conducting ISR from seabed to space and performing
joint all-domain operations anticipates a level of integration that provides
“commanders access to information to allow for simultaneous and sequential
operations using surprise and the rapid and continuous integration of capabili-
ties across all domains—to try to gain physical and psychological advantages
and influence and control over the operational environment” [1]. For example,
the Air and Space Force perspective is that, “In JADO, intelligence must de-
velop, maintain, and share an awareness of the operational environment that
spans geographic, functional, domain, classification, and organizational bound-
aries. The scope of awareness should include intelligence on ongoing opera-
tions, adversary forces, indications, and warnings (I&W), target information,
and account for military, political, and environmental considerations” [2]. The
high level of integration of information requires technology to provide broad
coverage, rapid access to collected information, and responsive control of the
collection network.
It is important to distinguish the terms that we use and explain the nu-
ances in terminology regarding the integration of diverse sources of intelligence:

107
108 Naval ISR Fusion Principles, Operations, and Technologies

Data fusion is a broad term that refers to the process of combining data or
information to estimate or predict entity states. The terms multisensory
and multisource data fusion, and information fusion all refer to technical
processes that perform data fusion. The term “process” can refer to human
activities (e.g., analysis) or to computational algorithms, but it is most ap-
plied to refer to the latter [3].
MultiINT refers to the automated processing of multiple shared sources
of raw or preprocessed intelligence to derive integrated intelligence prod-
ucts. The study of MultiINT processes is an interdisciplinary field seeking
to understand how integrating intelligence across the intelligence cycle
can vastly improve tactical and strategic decision-making. This is achieved
by using processes, algorithms, and systems to extract knowledge and in-
sights from data that ranges from being abundant to being sparse and
unstructured [4].
All-source intelligence analysis refers to the analytic process performed by
human analysts with access to all available intelligence sources. This term
was originally intended to distinguish all-source from single-source dis-
ciplines (e.g., SIGINT, GEOINT, HUMINT, and OSINT analysis) [5].

While intelligence in the maritime environment has always focused on

naval vessels, especially combatants, the expansive definition of the maritime
domain has expanded to include “all areas and things of, on, under, relating to,
adjacent to, or bordering on a sea, ocean, or other navigable waterway, includ-
ing all maritime-related activities, infrastructure, people, cargo, and vessels and
other conveyances” [6]. Complete all-domain naval intelligence preparation of
the operational environment and situation awareness should account for the
range of maritime battlespace dimensions that include [7]:

• Geographical features and meteorological and oceanographic character-

istics;
• Population demographics (ethnic groups, religious groups, age distribu-
tion, income groups, public health issues);
• Political and socioeconomic factors (economic system, political fac-
tions);
• Infrastructures, such as transportation and information systems;
• Rules of engagement (ROE) or legal restrictions on military operations
as specified in international treaties or agreements;
• All friendly and adversary conventional, unconventional, and paramili-
tary forces and their general capabilities and strategic objectives;
 Maritime MultiINT Fusion Processes 109

• Environmental conditions (earthquakes, volcanic activity, pollution,

naturally occurring diseases);
• Psychological characteristics of adversary decision-making;
• All locations of foreign embassies, nongovernmental organizations, and
private volunteer organizations.

In this chapter, we focus on MultiINT fusion as an automated process to

combine data across seabed-to-space sources to estimate the maritime threat
situation in all physical domains of operations, as well as in cyber, geopolitical,
and the other dimensions of the maritime domain. This chapter and the next
address solutions to the four challenges posed in Chapter 5, explaining how
automated and distributed data fusion enable distributed maritime operations.

6.1 The JDL Model for Organizing Naval ISR Fusion

The fundamental framework for describing the process for integrating data
from diverse sources is the JDL data fusion model. Developed in the early
1980s by a subpanel of the U.S. DoD Labs, the basic model has remained an
organizing structure to define the data fusion process:

…A process dealing with the association, correlation, and combination of

data and information from single and multiple sources to achieve refined
position and identity estimates, and complete and timely assessments of
situations and threats, and their significance. The process is characterized
by continuous refinements of its estimates and assessments, and the
evaluation of the need for additional sources, or modification of the process
itself, to achieve improved results [8].

This section provides a brief overview of the model related to naval ISR;
established texts provide a through treatment of the mature model [9]. The
most basic JDL model (Figure 1.6) represents the flow of analysis and, as an
abstract model, can refer to a process performed by humans, machines, or both.
It is similar to a fundamental model of human cognition that has the stages (or
levels) of sensation, perception of objects, perception of situations of objects
over time, and reasoning that construct meaning, followed by action. The Kah-
neman research in human cognition has distinguished between two modes, or
systems of thought: fast (intuitive, emotional) reaction, and slow (deliberative
and logical analysis); these feedback loops to response are indicated by the (1)
and (2) arrows in the model, respectively [10]. The fast loop includes immedi-
ate actions in response to a detection (e.g., new vessel appears in threatening po-
sition) and reactive tasking to identify the vessel. The full JDL loop includes the
110 Naval ISR Fusion Principles, Operations, and Technologies

complete deliberative reasoning through all levels of the process. The SHOR
(sense, hypothesize, orient, respond), OODA, TCPED (task, collect, process,
exploit, disseminate), and other C2 models are similar efforts to represent the
fundamental elements of observing, thinking, and response.
The benefit of the JDL model is that it provides an organizing scheme for
distinguishing the stages or levels of abstraction at which cognition, analysis,
or automated fusion is performed. The levels of processing (or, cognition) are
distinguishable, each with a definable role in moving from raw sensation to
perception, comprehension, meaning (relative to a mission goals), and intuitive
reaction and reasoned response.
From the basic model we can organize a more detailed naval ISR model
(Figure 6.1) that distinguishes the four most basic JDL levels of processing in
the maritime context [11].
Level 0. In some cases, it may be possible to combine raw data from mul-
tiple sensors prior to the detection in each individual sensor. This upstream fu-
sion can be particularly valuable when the signals in both sensors are weak and
even undetectable independently. The process minimizes information loss that
results from the detection process in each sensor that may reject weak signals;
the combination of correlated weak signals across both sensors may result in a
confident detection [12]. The process occurs upstream, and sensors must pass
the predetection data to the level 0 fusion processor.
Level 1. Object refinement performs the detection of objects, and refines
the estimate of state (location, identity, track behavior, etc.) and maintains cus-
tody of the object over time as sensing data allows. Sensor data must be cor-
related (measure the degree of similarity of two observations) and associated
(decision to assign observations across sensors to a single object). Once associ-
ated, the sensor data is used to refine the state and then predict the next state at
the next collection opportunity. In the maritime case, for example, a ship’s radar
track may be correlated, then associated with a commercial satellite’s AIS track
of the ship to identify the vessel. The specific processes performed here include:

• Alignment: First, the data from different sources must be aligned with
a common time and space reference system (as well as common scale of
confidence, probability, and other measurements).
• Correlation: Observations must be compared by some measure of simi-
larity or closeness in time, space, or attributes.
• Association: Once the correlation measurement between observations
meets a likelihood criterion for similarity, the observations may make an
association decision to assign the observations to the same target object.
• Tracking: The state of sensor observations over time may be computa-
tionally modeled to create a track—a state model of the object initiated
 Maritime MultiINT Fusion Processes
111

Figure 6.1 The JDL model as a closed-loop sensor control system.

112 Naval ISR Fusion Principles, Operations, and Technologies

and updated over time. The track model may be used to predict the
future position of an object for comparison (correlation and association)
with a future observation.
• Identification: Identification is the process to determine the type of a
target (e.g., a surface vessel), the classification (e.g., a merchant ship or
a guided missile destroyer model 55), and identity (e.g., hull number
047).

Level 2. Situation assessment considers the relations among objects (in

time, space, and behavior) to assess the meaning of any relations; this is a higher
level of abstraction and considers the events, activities, context, and timing fac-
tors to derive a situation estimate. In the maritime case, this will assess the ag-
gregate behavior of adversary combatant vessels, their time-from-port, fuel, and
readiness status, coordinated behavior (maneuver, active sensors, communica-
tions, etc.), predicted mission, and intent. Even higher-level context must be
considered, including geopolitical context, weather and sea state, and windows
of commercial satellite coverage [13].
Level 3. Impact assessment is where the estimated situation is compared
to the fleet mission and the effects of the current situation (e.g., Blue fleet and
Red fleet) are estimated. The potential vulnerabilities of Blue (e.g., exposure to
Red sensors and weapons) as well and the opportunities for Blue (e.g., covert,
undetected sensing of Red) are projected for the commander [14].
Level 4. Process refinement performs the ISR resource management to
orchestrate the network of sensors to close intelligence gaps and focus tracking
to maintain custody on major threats (or to resolve key unknowns). This is a
feedback level, adapting collection by coordinating all ISR sensors to optimize a
sensing objective function (maximum custody of key targets, maximum search
coverage, maximum track accuracy on key threats, etc.). While this feedback
is essential for an adaptive sensing system on a single platform, it is also es-
sential for a distributed system where distributed operations require control of
widely distributed sensors (seabed to space). In a distributed environment level
4 is required to: (1) orchestrate across all sensors to assure coverage (for search
and surveillance) and focus (for targets that must remain in custody), and (2)
adapt the computation and network communication to maintain processing
and bandwidth allocations based on the situation.
The JDL model figure only represents the abstract levels of a single fusion
node that combines multiple sources, but we must recognize that the distrib-
uted naval fleet has many interconnected fusion nodes that perform different
levels of the fusion process on different target types (air, surface, and subsur-
face). Each node is unique to its mission and its access to sources; and it may
be reporting fused information to other fusion nodes to contribute their fusion
 Maritime MultiINT Fusion Processes 113

process. To avoid over-reporting (double-counting), the fused information must

maintain a degree of metadata to enable appropriate fusion across the network:

• Data pedigree is the description of data source and attributes of collec-

tion and fusion process (generally maintained in metadata that uniquely
defines the data itself and provides a traceable path to its original col-
lection).
• Data provenance is a broader term that often includes pedigree as well as
the process (or network of fusion nodes) by which it arrived in an intel-
ligence or fire control database, and processing performed along the way.
Aspects of provenance also include data quality and attribution (source,
sensor, platform, classification, etc.) Some degree of provenance can pre-
vent double-counting and over-processing combined information.

Figure 6.2 is a basic example to illustrate just one possible fusion archi-
tecture that allocates processing at different JDL levels across different naval
platforms. Note that we only included one of each type of platform to simplify;
a fleet architecture will include scores of platforms and hundreds of potential
fusion nodes. The figure points out three implementation issues that are impor-
tant for distributed naval implementations:

• Central fusion: Major nodes will perform all levels of fusion and distrib-
ute results (common operating picture) to other nodes. It is critical to
retain pedigree of data and fused results to avoid multiple counting of
objects, events, or activities.
• Distributed fusion process distribution: Fusion processes may occur at all
nodes in the naval network (sensor nodes, relay nodes, and C2 nodes).
Even a sensor node, for example, may receive data from other sensors
(e.g., tips to focus on a target) and may perform some form of fusion
for its own functions (e.g., to evade threats or to adapt processing and
sensing). We discuss this further in Chapter 7.
• All-domain fusion: While the JDL model was initially applied to physical
objects (ships, submarines, and maritime aircraft) it has also been ap-
plied to represent the processing of sensor and source data on nonphysi-
cal objects of interest to naval intelligence: signal contents, cyber-attack
processes, social media messages, social attitudes of ashore populations,
and networks of adversarial naval organizations. Naval ISR must con-
sider the integration of all these domains to assess the overall situation
and understand a coordinate multidomain activity.
114 Naval ISR Fusion Principles, Operations, and Technologies

Figure 6.2 JDL structure for elements of a maritime fusion network.

In the next sections we expand on the operations performed at each of the

levels in the maritime context.

6.2 Maritime Object and Situation Assessment Levels 1 and 2

The result of an automated level 1 process (also called object-based production
(OBP)) is a dynamic data base that contains all objects in a theater of operations:
objects, unique identifiers, their identities if established, the dynamic tracks if
established, links to prior activities of an identified object (port calls, patterns
of behavior, etc.), basic data on the object (e.g., basic ship characteristics and
capabilities), and additional assigned data (e.g., the object relative priority).
We illustrate the implementation of a representative level 1, 2, and 4 pro-
cess for a maritime situation assessment mission in Figure 6.3 where a sequence
of commercial satellite sensor observations (e.g., SAR and EO, as described in
 Maritime MultiINT Fusion Processes 115

Figure 6.3 Maritime situation assessment and collection planning process.

Chapter 9) are collected and processed to derive a maritime situation. In ad-

dition, we may have access to commercial AIS data that reports the identity
and location of vessels to compare with the EO and IR imagery detections. We
adopt the key terms (for a thorough overview see [13]):

• Activity refers to a recognizable movement, conducted by a single entity

that is an indicator or has a specific meaning when viewed within a rel-
evant context [15];
• Situation refers to a set of environmental elements and events or activi-
ties with respect to time or space [16];
• Situation detection refers to the process of perceiving, identifying, and
declaring specified situations;
• Situation awareness refers to the process of perceiving the environmental
elements and events or activities with respect to time or space, the com-
prehension of their meaning, and the projection of their future status.
(Inferred from [16].)

This maritime implementation is a model-based approach to fusion be-

cause it applies reasoning that compares sensor and source evidence to known
threat models (e.g., ship object models, ship and aircraft behavior models, and
ship and aircraft electronic signature models). The process attempts to instanti-
ate models as target hypotheses that can then be developed by seeking addition-
al supporting or refuting evidence for the hypothesis, by requesting additional
ISR collection. The model-based fusion and data-driven collection approaches
were recommended by a joint DoD-Intelligence Community Science Board for
these kinds of ISR challenges [17].
The automated process proceeds in the sequence that follows (Figure 6.3):
116 Naval ISR Fusion Principles, Operations, and Technologies

1. Level 0: Detect objects upstream. In some cases, it may be possible to

combine raw data from multiple sensors prior to the detection in each
individual sensor. This upstream fusion can be particularly valuable
when the signals in both sensors are weak and even undetectable inde-
pendently. The process minimizes information loss that results from
the detection process in each sensor that may reject weak signals; the
combination of correlated weak signals across both sensors may result
in a confident detection [12]. The process occurs upstream, and sen-
sors must pass the predetection data to the fusion processor. In this
case, the raw EO and SAR data would have to be processed together
to detect weak image targets.
2. Level 1: Detect and refine maritime objects. Radar and EO imagery are
processed to detect and extract objects (vessels) and their locations
(all imagery is first orthorectified to remove geometric distortions and
reference to map coordinates). At this stage the objects may be cor-
related to AIS transponder reports to identify those vessels reporting
their identity. The extracted radar and EO image chips (small images
that just contain the vessel object) are passed to a target classifier (e.g.,
a neural network classifier) to determine the object class (merchant,
naval combatant, fishing) and type (e.g., cruiser, destroyer, frigate).
This process is referred to as OBP where POSIT objects are extracted
and placed in an object store for: (1) subsequent tracking processes,
(2) tipping to cue up other sensors (or cross-cueing) to identify an ob-
ject, and (3) development of the pattern of life of the objects or clus-
ters of objects. OBP is the process of organizing intelligence around
objects of interest as a means of organizing information for analysis.
OBP methods provide digital intelligence reports to reflect all of the
information on objects (real-world things and events) and to deliver
assembled reports that include all relevant information [18]. All ob-
servations are managed (indexed, described, and organized) in a com-
mon structure and are based on the object observed—this process is
called structured observation management and allows MultiINT sharing
across organizations [19].
3. Level 2, Step 1: Detect activities. Activities are characterized by objects
(vessels) moving and interacting with other objects and significant
locations. Activities obviously occur over a specified period of time.
Thus, detecting an activity involves looking for correlations in the
relationship of these objects and locations in space and time. Several
classes of algorithms can be applied to perform the detection with this
data, ranging from track template matching to learned neural nets.
The relative effectiveness of these techniques will depend on the qual-
 Maritime MultiINT Fusion Processes 117

ity and quantity of examples available for the activities to be detected.

Some machine learning techniques can detect coordinated actions
that may indicate an activity that does not match a previously learned
and labeled pattern. Research has demonstrated the performance of
processes to detect and identify ship tracks based on the correlation of
AIS tracks, satellite SAR ship detections, and medium spatial resolu-
tion satellite imagery [20].
4. Level 2, Step 2: Assess situation. Once an activity or set of activities are
recognized, the time series of activities, and any correlation between
their activities and contextual factors (weather, tactical, etc.) are used
to assess the meaning of the situation. This assessment can be in the
form of a recognized tactical maneuver, a maritime activity (e.g., at-sea
refueling), a projected future (loitering in preparation for convoy), or
other maritime situations. The importance of the activity can then be
assessed. More importantly for our purposes, the degree of certainty
about the situation and the projection of the situation at future times
can be evaluated. This phase can include deeper automated and hu-
man-machine analysis using additional contextual data. For example,
one situation assessment process detected suspicious vessel loitering
behaviors to the graph network of shipping companies, their address-
es, owners, and the countries in which they are flagged. The methods
sought to identify gray (nontransparent) shipping networks that could
support state-directed hybrid maritime warfare [21]. The additional
context of the shipping ownership helped score the most likely en-
ablers of hybrid activities. The networks could be expanded to include
historical behaviors (e.g., suspicious loitering, commercial vessels co-
traveling with military vessels, or being fitted at military shipyards)
of vessels. If the development of the future situation is uncertain, but
it can be asserted that specific data elements could resolve or reduce
the uncertainty, then a collection task may be created to address the
uncertainty.
5. Level 4: Plan/schedule next collection. If collection of certain data can
disambiguate a developing situation, then there are two basic ques-
tions that need to be answered in order to plan the appropriate tasking
actions by available sensors. The first question is “What objects or
activities need to be ascertained, and what sensors can supply that in-
formation?” The second question is, “Over what future time horizon
is this data likely to be available or useful and when is it needed to im-
pact the situation?” Additionally, we must consider the relative value
of the many targets we wish to observe. (By value, we mean a measure
of mission importance. Often value is used to describe a candidate col-
118 Naval ISR Fusion Principles, Operations, and Technologies

lection and may depend on target importance as well as type of sensor

data to be collected, time, and location.) Once these questions are an-
swered we are left with a straightforward (though computationally dif-
ficult) optimization problem. The scheduling process must consider
a large number of alternative collection options and sequences, then
choose among alternatives, based on maximizing a value-based objec-
tive function. (We discuss this further in Section 6.4.) In some cases,
trained collection managers can perform this function, but for many
cases (such as satellite collection) the speed and number of options to
be considered dictate the need for automation. For modest numbers
of sensors over a short timeframe, techniques such as mixed-integer
linear programming are appropriate. For very large numbers of sensor
opportunities and/or targets, more powerful techniques are required.
Furthermore, it is likely that a truly optimum solution will not be
obtainable in time to task sensors and that a satisfactory solution must
be sought instead.

The preceding example illustrated the processes for levels 0, 1, and 2 for
the traditional naval maritime domain awareness mission, focusing on physical
objects of military interest. It is necessary to point out that the fusion process
is also applied to discover, understand, and model the dynamics of entities that
are nonphysical (social networks, populations, economies, etc.). Nonphysical
entities are the objects (targets) of information operations and special opera-
tions focused on organizations whose dynamics do not follow the kinematic
behaviors of the physical world. Navies have ISR tasks that require monitoring
social influence in ashore populations, civilian ashore social networks, criminal
human networks, adversary command organizations and their human decision-
making, and more. Activity based intelligence (ABI) is a MultiINT discipline
characterized by analysis of human activity and transactional data to resolve un-
knowns using contextual knowledge of activities and transactions. The analytic
focus is on entities (people and their organizations), their relationships (net-
works of entities), their activities, and patterns of behavior. ABI is contrasted
with traditional target-based intelligence because of its focus on discovering the
relationships and behavior (activities) of networks (target systems) [22].
The contrast in the JDL levels between the physical and nonphysical tar-
get systems (Figure 6.4) shows the mapping between JDL target-based fusion
notions and nonphysical networks that are represented as graphs where objects
are referred to as entities, groups or units are referred to as networks or systems,
and target tracking is more generally referred to as behavior (kinematic track-
ing is replaced by social dynamics or network dynamics and modeled appro-
priately). The general term for a generic person, place, or thing is object, as we
 Maritime MultiINT Fusion Processes 119

Figure 6.4 Applying the JDL fusion model to nonphysical systems.

have used throughout. A specific named or labeled object (e.g., John A., Farris
Building, homemade explosive) becomes an entity, in some parlance.
The steps in Figure 6.4 applied to ISR of foreign ashore terrorist networks
resolves the individual human objects (generally called actors because they are
socially interacting humans) at level 1, then discovers the relationships (com-
munications, financial, influence, etc.) between actors to form networks at level
2. The structure can be studied by social network analysis (SNA) to derive the
importance and influence of the individual actors in the net; the timeline of
events of the network helps to explore behavior, patterns of life, and potential
future courses of action at level 3. Unique to complex human systems, the
feedback (level 4) actions not only include sensor management to refine the
network, but the use of probing actions to influence the complex adaptive net-
work to reveal information. Probes include any action designed to produce a
response that reveals new properties of the network. (The analogy in dynamic
systems is the injection of an impulse to measure the output impulse response
function that reveals properties of the system.) For example, the analyst refining
a terrorist network that is messaging on social media may send a set of mes-
sages designed to elicit a response to determine how (and who in the network)
responds to a meme or concept.
Example sources and analytic targets of nonphysical (or human) systems
of interest in the maritime domain include:

• Fishing fleets. Source: radio traffic in fishing fleet. Analytic targets: radio
network, legitimate and illegitimate activities, communicating parties;
120 Naval ISR Fusion Principles, Operations, and Technologies

• Ashore terrorist organizations. Sources: social media, ashore HUMINT.

Analytic targets: terrorist groups, actors, major influences, target audi-
ences of public messaging;
• Foreign naval units in area. Sources: radio traffic, ashore naval intelli-
gence, public port departure data. Analytic targets: crew, organizational
network across the fleet;
• Ashore Ports (prearrival). Sources: public broadcast, news, social media,
PAI, HUMINT. Analytic targets: threat and criminal organizations and
individuals, threat activities.

6.3 Maritime Impact or Threat Assessment Level 3

The definition of impact assessment introduced earlier in this chapter addressed
the need to identify the effect of the current (and projected) situations on the
vessel, force package, or fleet’s mission [13]. This effect, or mission impact,
includes both opportunities and threats. In the first chapter, we introduced
the wide range of naval domains (Section 1.1), missions (Section 1.2), and
operations (Section 1.2); in each there are unique contexts, dispositions, spatial
geometries, activities, and events that indicate opportunities to act, or threats to
avoid, deter, or defend.
The distinctions between threats and opportunities in the maritime con-
text are illustrated for the maritime superiority mission (Table 6.1) with specific
examples of the kinds of situations that naval officers are trained to recognize,
and that automated Level 3 fusion processes seek to recognize and adapt sensors
to observe. These kinds of situations include, for example:

• Maneuvering to certain relative geometries of vessels or aircraft (manned

and unmanned);
• Positioning of vessels or aircraft to critical distance between vessels rela-
tive to weapon ranges;
• Timing of events and sequences of activities;
• Coordinated behaviors of assets (subsurface, surface, air, and overhead
surveillance) including loitering, merging, co-traveling, maneuvering to
firing positions or protective positions, deceptive behaviors, massing of
forces, and more;
• Electromagnetic activities such as going dark with EMCON, illumina-
tion of vessels with fire control sensors (e.g., high pulse repetition fre-
quency radar), coordinated electronic activities, jamming, spoofing or
other deceptive actions, and so on.
 Maritime MultiINT Fusion Processes 121

Table 6.1
Maritime Superiority Opportunities and Threats
Area Maritime Opportunity (O) Maritime Threat (T)
Objective Offense: Degrade, disrupt, deceive, deny, Defense: Protect, defend, recover own
and exploit enemy’s IE to introduce fog; IE, and increase the enemy’s friction by
mass distributed forces decisively engage creating uncertainty.
the enemy’s critical vulnerabilities to
degrade and remove the enemy’s center
of gravity.
Example Opportunity situations for Blue: Threatening situations to Blue:
T/O ·· Red vessel or aircraft position, ·· Red vessels have successfully
Behaviors electronic state, sensor positioning, maneuvered (electronic, cyber, physical)
communication ability, or other such that Blue loses custody;
condition degrades its situation ·· Red forces (surface, air) are massing
awareness; and maneuvering in the direction of Blue
·· Red sensor gap enables a deceptive forces;
maneuver or EM action; ·· Red platforms are surveilling distributed
·· Red network degradation, even if Blue forces to the degree they have
temporary; near-full custody of Blue;
·· Red vessel positions are such that a ·· Red raises intensity of offensive
sensor ambush can be induced by Blue cyber operations coordinated with
actions; or preceding threatening physical
·· Red forces are dispersed to a high behaviors;
degree of vulnerability to electronic ·· Red forces exhibit coordinated denial
attack to disrupt networking. (e.g., go AIS dark) and deceptive (e.g.,
AIS spoofing) behaviors.

The role of level 3 is critical to provide warning to C2 officers and systems

of situations that require attention and action. Level 3 simulations that explore
future courses of action may be integrated with C2 systems to allow operations
planners to explore future courses of action and potential outcomes. The cur-
rent situation simulation (provided by the ISR model) can be extended forward
in time applying alternative Red and Blue COAs to examine potential effects
and outcomes [23]. In Chapter 8, we introduce further potential contributions
from AI technologies to aid decision-makers with broader context, situation
recognition, and COA simulation [24].
Automated level 3 processes implement in a computational form, the
maritime version of the Joint Intelligence Preparation of the Operational Envi-
ronment (JIPOE) process described in DoD publication JP 2-10.3 [25]. JIPOE
is the analytical process used by joint intelligence organizations to produce in-
telligence assessments, estimates, and plans in support of the joint force com-
mander’s decision-making process. JIPOE is a continuous process that includes
defining the current operational environment, evaluating the adversary, and
determining and describing the adversary’s potential COAs to identify opportu-
nities and threats. The four-step analytic process (1) describes the operating en-
vironment and the kinds of effects that are relevant to a commander’s mission,
(2) describes the impact of the operational environment from a comprehensive
122 Naval ISR Fusion Principles, Operations, and Technologies

systems-level viewpoint, including friendly and adversary forces as well as geo-

spatial considerations, (3) evaluates the adversary in the context of the environ-
ment, and then (4) anticipates the likely and threatening COAs the adversary
may take (Table 6.2).

6.4 Maritime Distributed Resource Allocation and Orchestration

The allocation of maritime ISR sensors is performed at several levels of consid-
eration (Table 6.3), beginning with a strategic assessment across the fleets and
action groups, then at the operational fleet level to assign organic platforms

Table 6.2
Computational Steps in the Level 3 JIPOE Process
Representative Level 3 Fusion Computational Modeling Activities (with JP
JIPOE Step 2-10.3 References)
Step 1: 1. Model the commander’s intent for the JIPOE effort (JP2-10.3; II-4).
Define the 2. Identify and model key effects relevant to the commander’s intent and the
Operational systems within the environment that must be represented.
Environment 3. Maintain a model library of the environment (JP2-10.3; II-5); identify the
appropriate models and data required to instantiate component models and meta
model for the environment (geospatial, weather, naval behavior, etc.).
4. Compose a baseline model, calibrate and validate behavior compared to prior
months as appropriate for the environment and current situation.
Step 2: 1. Update by ISR the current situation in the operational environment (JP2-10.3;
Describe the II-11e Information Domain; II-12 Systems Perspective). For human social cultural
Effects of the behavior issues, evaluate policy leaders, political parties, key population groups,
Operational spoilers (e.g., terrorist groups), and their relations, positions, and relative political-
Environment social power. For the operational context, assess economics, corruption, security and
terrorist trends, and media influences.
2. Accept and update information from current intel systems to establish model
initial conditions and parameters.
3. Consider future outlook and impact of Op environment on adversary (Red) and
Blue COAs. Using baseline simulation, project effects if no Red actions are taken
and explain the dynamics of the environment (JP2-10.3; II-13 Systems Perspective).
Step 3: 1. Review the modeled information about the adversaries and their relationships
Evaluate the with other groups (JP2-10.3; II-14-16); identify uncertainties and sensitivities.
Adversary 2. Identify sources of power and relative effects of alternative positions (JP2-10.3;
II-17).
3. Identify relevant centers of gravity (JP2-10.3; II-18) across the Red and Blue
systems.
Step 4: 1. Perform analytic simulations to identify potential adversary objectives and
Determine endstates (JP2-10.3; II-19-20).
Adversary 2. Perform a range of exploratory simulations to evaluate potential effects of
COAs adversary activities (e.g., maritime expeditionary campaigns, all-domain coordinated
campaigns, political action, and media campaigns) to achieve their desired Red
objectives and endstates (JP2-10.3; II-21).
3. Develop and evaluate the effectiveness of alternative adversary COAs (e.g., media
campaigns, adverse influence-economic actions, maritime piracy, and maritime
expeditionary campaigns).
 Maritime MultiINT Fusion Processes 123

Table 6.3
Levels of Sensor Systems Allocation
Level Description and References
1. Strategic ISR force Allocation of the ISR platforms across fleets to prioritize coverage
management. of multiple operating areas [26].
2. Operational force ISR Allocation of ISR platforms to multiple AOIs within a fleet [27].
allocation.
3. Tactical dynamic ISR sensor Dynamic allocation of sensors-to-targets on individual ISR
management. platforms in near real-time in a networked sensor environment [28].

across areas of interest (AOIs). Each allocation is based on an analytic process

that must place a value on maritime activities in each area, and optimally al-
locate sensors platforms to activities. In this section, we consider the real-time
dynamic allocation of sensors to search areas, surveillance areas, targets, target
tracks, and target groups in the fusion level 4 process.
The fourth fusion level manages the sensing capability of the system; in a
maritime system, we refer to the network of sensors available to the fleet ranging
from the seabed to space. In an ideal architecture, a network to all sensing re-
sources could be accessible to the fleet and fully controllable to optimally sense
the maritime environment, focus on threats, and adaptively manage the sensor
platform or network to concurrently perform search, reconnaissance, surveil-
lance, identification (e.g., dwell a sensor on an individual target, or interrogate
a target for identification), and tracking to support engagement.
The management of the network to coordinate all sensors as an adap-
tive sensing network is referred to as sensor or collection orchestration, and as a
conductor harmonizes a multitude of instruments to produce harmony, this
process seeks to optimize a multitude of sensors to produce feasible assignments
and an optimum collection schedule. We consider two categories of required
coordination control to maintain custody of many maritime targets for the dis-
persed DMO fleet (Table 6.4) [29].

• Sensor or collection orchestration is the structure of applications required

to manage the sensor collection and network paths over time to proceed
from search, detection, track-ID, and fire control. This fusion level 4
process requires the automation of the distributed network of MultiINT
collection and processing elements to schedule dynamic collection on
emergent events as well as preplanned (standing) collection needs in an
optimal way.
• Network orchestration: The network must adapt to vary coverage, focus,
and priority as the system changes the importance of different missions
(ocean search, ship-track, reacquire, and engagement), areas, and targets.
The network must allocate bandwidth as phases of the mission change
124 Naval ISR Fusion Principles, Operations, and Technologies

Table 6.4
Orchestration Architecture Elements
Sensor Orchestration Network Orchestration
MultiINT Sensor Planning Based on Target Value Route Planning Based on Message Priority
Target Sensor Aggregate Value Network Routing Message Priority Based
Valuation Collection Optimization Prediction on Current Projected
Feasibility ISRT State
Prediction
·· Coordination of multiple sensors to sustain ·· Network instantaneous performance
continuity of surface, subsurface, or air target management priority message stream
custody; assurance (e.g., weapon inflight; target
·· Prediction and valuation of each feasible sensor in-track).
collection;
·· Assignment of values for priority targets-in-track.

over time. Of course, engagement operations must have priority during

fire control activities that require sensors to provide frequent updates
throughout missile flight.

Table 6.4 illustrates the specific functions in each category of orchestra-

tion; both categories must be coordinated to operate the network and sensors
to optimize an objective function. Then objective function must consider the
balance between several alternatives:

1. Objective 1: Maintain custody of the maximum number of targets

identified by the fleet as important and within DMO operating areas.
2. Objective 2: Maintain custody to achieve the highest aggregate target
value, where each target has a defined value (e.g., an aircraft carrier has
a higher value than a frigate).
3. Objective 3: Search an area to locate and acquire a threat vessel by
maximizing the greatest coverage area in the shortest time.
4. Objective 4: Maintain network flow to minimize latency-inducing
congestion, particularly during high-dynamic periods (the heat of na-
val battle).

When optimizing sensor orchestration for a value there are two contribu-
tors to the value of a candidate target (or area) for collection.

• Target value is the importance of a target, object, or activity (or an area

when searching) that requires collection. High importance targets have
a higher value. This is related to the intrinsic intelligence value of the
knowledge about that target.
 Maritime MultiINT Fusion Processes 125

• Opportunity value is the expected value of the collection on the target

under predicted collection conditions (range, SNR, etc.). Of course,
poorly observed collections (due to weather, obscuration, long range,
etc.) have lower values.

The orchestration process then applies an algorithm to assign an overall

value (the collection value) to each collection request item (e.g., maritime tar-
get, or ocean search area) based on the combination of its target and opportu-
nity values. The expected value must be based on a quantitative representation
of the commander’s intelligence requirements (CIRs).
The overall values of all targets (at a point or window of time) can be rank-
ordered and optimized to achieve the highest overall value. Several approaches
to optimization may be applied to perform the optimization, including:

• Linear programming maximizes (or minimizes) a linear objective func-

tion subject to one or more constraints, where all the constraints are
linear, and all of the search variables are continuous variables. These
constraints are generally too restrictive for sensor assignment problems.
• Mixed-integer programming adds one additional condition that at least
one of the variables can only take on integer values. This method is most
applicable to many assignment problems that are typical of sensor col-
lection and resource allocation.
• Simulated annealing is a probabilistic technique for approximating the
global optimum of a nonlinear objective function.
• Genetic search methods that apply heuristics to search large spaces (ob-
jective functions) either constrained or unconstrained, by expanding a
search, choosing the best results, pruning unpromising results, and re-
peating the search for a global optimum.

The purpose of the optimization is to use the collection resources avail-

able most efficiently and to dynamically focus attention on the most critical
maritime targets. The effect is a dynamic control of the network and collection
resources. But a widely distributed network of sensors is far from ideal and must
address several challenges:

• Control: Sensors vary widely in their ability to control their sensing

modes, pointing, and access times. In addition, some sensors operate
on platforms (spacecraft, aircraft, ships, or submersibles) that require
the coordination of platform steering to position the sensor to have ac-
cess (generally line-of-sight) to an area of target. Other sensors are fixed
126 Naval ISR Fusion Principles, Operations, and Technologies

(seabed sensors) or floating (surface wave-glider sensors) or governed by

regular physical laws (Kepler’s laws for spacecraft orbits) and only revisit
areas or targets on a schedule.
• Sensor coverage and target access: The field of regard, instantaneous field
of view, and sensor platform motion determine the feasibility of a given
sensor having access to observe a target, and the duration of the observa-
tion (the sensor dwell on a target). Coordination of sensors to observe
multiple targets requires the ability to predict the feasibility of each sen-
sor to have access to each target.
• Network latency: The delays in networks limit the speed-of-control of
sensors as well as the speed-of-response in sensor reporting. This delay
can limit the performance of a networked sensor as the latency introduc-
es a phase delay in closed-loop response of the sensor; this can severely
limit the sensor’s ability to track a ship, for example. Furthermore, the
delays from sensing to fusion across such a wide range of sensors can di-
minish the ability to perform some joint sensing modes (e.g., identifica-
tion of a target by the features provided by two sensors at a coordinated
time). This is not a large factor in single platform fusion, such as fusing
the SAR, EO, and RF sensors on a maritime patrol aircraft, but it is a
significant factor on widely distributed sensors.

6.5 Conclusion
The subject of distributed sensor management is complex and the scholarly
materials on algorithms for determining feasible solutions, selecting objective
functions, and applying optimization methods is vast. Greater depth in this
subject is beyond the scope and level of this book, but we recommend the fol-
lowing keys sources that address this challenge from different perspectives:

• Hintz, K. J., Sensor Management in ISR, Norwood, MA: Artech House,

2020.
• Hero, A. O., and D. Cochran, “Sensor Management: Past, Present, and
Future,” IEEE Sensors Journal, Vol. 11, No. 12, December 2011.
• Hero, A. O., et al., Foundations and Applications of Sensor Management,
Springer, 2007.
• Johnson, B. W., and J. M. Green, “Naval Network-Centric Sensor Re-
source Management,” Naval Postgraduate School, April 2002.
• Ng, G. W., and K. H. Ng, “Sensor Management—What, Why And
How,” Information Fusion, Vol. 1, No. 2, 2000, pp. 67–75.
 Maritime MultiINT Fusion Processes 127

• For application to a maritime surveillance problem involving the alloca-

tion of multiple heterogeneous assets over a large area of responsibility
to detect multiple drug smugglers, see Zhang, L., “Context-Aware Dy-
namic Asset Allocation for Maritime Surveillance Operations,” Journal
of Advances in Information Fusion, Vol. 15, No. 1, June 2020.

This chapter introduced the key functions, as if the fusion functions are
performed in a centralized process; this may indeed be true on some platforms.
But across the DMO fleet, there may be many fusion processors that share data
and products across the fleet network. This network sharing and network fu-
sion poses new changes. In the next chapter, we move to describe approaches to
implement these fusion functions across networks where the fusion functional-
ity is distributed and shared.

Endnotes
[1] Congressional Research Service, Joint All-Domain Command and Control (JADC2),
IF11493, Version 16, January 21, 2022.
[2] Air Force Doctrine Publication 3-99 and Space Force Doctrine Publication 3-99, Depart-
ment of the Air Force Role in Joint All-Domain Operations, November 19, 2021, p. 13.
[3] Steinberg, A. N., C. L. Bowman, and F. E., White, “Revisions to the JDL Data Fusion
Model,” Proc. Third NATO/IRIS Conference, Quebec City, Canada, 1998.
[4] Center for MultiINT Studies, Naval Postgraduate School, https://nps.edu/web/cmis.
[5] Fingar, T., “A Guide to All-Source Analysis,” The Intelligencer, Journal of U.S. Intelligence
Studies, Vol. 19, No. 1, Winter/Spring 2012.
[6] IMO definitions. See definition in “Executive Order on Promoting American Seafood
Competitiveness and Economic Growth,” White House, May 7, 2020, https://www.
wpcouncil.org/wp-content/uploads/2021/05/Executive-Order-on-Promoting-American-
Seafood-Competitiveness-and-Economic-Growth-_-The-White-House.pdf.
[7] Joint Staff, Joint Tactics, Techniques, and Procedures for Joint Intelligence Preparation of
the Battlespace, Joint Publication 2-01.3, May 2020, pp. 24, II-4.5. Earlier doctrine in
the period from 2003 to 2020 included distinct domains (air, land, maritime, and space),
the IE (which includes cyberspace), and political, military, economic, social, information,
and infrastructure (PMESII) systems and subsystems. Joint Staff, Joint Intelligence Prepa-
ration of the Operational Environment Joint Publication 2-01. 3, May 21, 2014, p. 1-1.
For a critique of the PMESII elements, see Ducote, B. M., “Challenging the Application
of PMESII-PT in a Complex Environment,” Fort Leavenworth, Kansas: School of Ad-
vanced Military Studies United States Army Command and General Staff College, April
26, 2010.
[8] White, F. E., Jr., Data Fusion Lexicon, Joint Directors of Laboratories, Technical Panel
for C3, Data Fusion Sub-Panel, Naval Ocean Systems Center, San Diego, 1987. See also
128 Naval ISR Fusion Principles, Operations, and Technologies

White, F. E., Jr., “A Model for Data Fusion,” Proc. 1st National Symposium on Sensor Fu-
sion, Vol. 2, 1988.
[9] The following texts provide more depth on the data fusion model and process:
Handbook of Multisensor Data Fusion: Theory and Practice, Second Edition, M. Liggins, II,
D. Hall, and J. Llinas (eds.), Electrical Engineering & Applied Signal Processing Series,
CRC Press, 2008.
Hall, D., and S. McMullen, Mathematical Techniques in Multisensor Fusion, Norwood,
MA: Artech House, 2004.
Antony, R., Principles of Data Fusion Automation, Norwood, MA: Artech House, 1995.
Waltz, E., and J. Llinas, Multisensor Data Fusion, Norwood, MA: Artech House, 1990.
[10] We introduced this comparison earlier in Section 1.4. See Kahneman, D., Thinking, Fast
and Slow, Farrar, Straus and Giroux, 2011.
[11] In this brief introduction we include the four basic levels. Level 0 was added for
predetection fusion, or fusion at level of raw signals across multiple sensors. For additional
changes and considerations, see [3] and Chapter 3 in Handbook of Multisensor Data Fusion:
Theory and Practice, Second Edition, M. Liggins, II, D. Hall, and J. Llinas (eds.), Electrical
Engineering & Applied Signal Processing Series, CRC Press, 2008.
[12] Newman, A. J., and G. E. Mitzel, “Upstream Data Fusion: History, Technical Overview,
and Applications to Critical Challenges,” Johns Hopkins APL Technical Digest, Vol. 31, No.
3, 2013, pp. 215–233.
[13] Alan N. Steinberg, “Foundations of Situation and Threat Assessment,” in: Handbook of
Multisensor Data Fusion: Theory and Practice, Second Ed., M. Liggins II, D. Hall, and
James Llinas (eds), Chapter 18, 2009.
[14] For an example of a Level 3 implementation for a complex land scenario, see Chen, G., et
al., “Game Theoretic Approach to Threat Prediction and Situation Awareness,” Journal of
Advances in Information Fusion, Vol. 2, No. 1, June 2007, pp. 35–48.
[15] GEOINT CONOP 2022, National Geospatial Intelligence Agency, April 1, 2016.
[16] Endsley, M. R., “Toward a Theory of Situation Awareness in Dynamic Systems,” Human
Factors Journal 37 (1): 32–64.
[17] Report of the Joint Defense Science Board-Intelligence Science Board Task Force on
Integrating Sensor-Collected Intelligence Office of the Under Secretary of Defense for
Acquisition, Technology, and Logistics, November 2008, pp. 49–53.
[18] Johnston, C., “Modernizing Defense Intelligence: Object Based Production and Activity
Based Intelligence,” Briefing, June 27, 2013.
[19] The term structured observation (or systemic observation) is a method of collecting and
recording data used in numerous scientific disciplines. The structured procedures guide
the method of observation and encoding of the observations for subsequent study. The
terminology is adopted to refer to methods of structuring information from geospatial
analysis for entry into object-based production systems.
[20] See the example that uses AIS and satellite imagery: Štepec, D., T. Martincic, and D.
Skoaj, “Automated System for Ship Detection from Medium Resolution Satellite Optical
Imagery,” Proc. IEEE Oceans 2019, October 1, 2019.
 Maritime MultiINT Fusion Processes 129

[21] Callaghan, C., R. Schroeder, and W. Porter, “Mapping Gray Maritime Networks for
Hybrid Warfare,” COMSEC, July 1, 2020, https://cimsec.org/mapping-gray-maritime-
networks-for-hybrid-warfare/.
[22] Biltgen, P. and S. Ryan, Activity Based Intelligence, Norwood, MA: Artech House, 2015,
and Antony, R., Data Fusion Support to Activity-Based Intelligence, Norwood, MA: Artech
House, 2015. See also the GEOINT perspective in Long, L., “Activity Based Intelligence:
Understanding the Unknown,” The Intelligencer: Journal of U.S. Intelligence Studies, Vol.
20, No. 2, Fall/Winter 2013, pp. 7–15.
[23] The integration of intelligence simulations and operations planning is introduced in
Waltz, E., Quantitative Intelligence Analysis: Applied Analytic Models, Simulations, and
Games, Rowman Littlefield, 2014.
[24] In addition to AI technologies, AI-enabled commercial data sources can contribute to
assessment. For example, geopolitical context may be aided by the open-source Global
Database of Events, Language, and Tone (GDELT) that monitors “the world’s broadcast,
print, and web news from nearly every corner of every country in over 100 languages and
identifies the people, locations, organizations, themes, sources, emotions, counts, quotes,
images and events.” See The GDELT Project, https://www.gdeltproject.org/.
[25] Joint Chiefs of Staff, Joint Publications (JP) 2-01.3, Joint Intelligence Preparation of the
Operational Environment, Joint Chiefs of Staff, Washington D.C., May 21, 2014.
[26] For an explanation of the high-level process to allocate ISR assets at the platform level,
see Carrillo, G. I., “Optimization Case Study: ISR Allocation in the Global Force
Management Process,” Naval Postgraduate School, September 2016.
[27] Ranjeev Mittu et al., “Optimization of ISR Platforms for Improved Collection in Maritime
Environments,” Washington D.C.: Naval Research Laboratory, August 2009.
[28] Johnson, B. W., and Green, J. M., Naval Network-Centric Sensor Resource Management,
2002, http://hdl.handle.net/10945/37809.
[29] Scrofani, J. W., and Miller, D. L., “All-Domain Sensor Network Orchestration from
Seabed-to-Space,” Monterey, California: Naval Postgraduate School, http://hdl.handle.
net/10945/69752.
7
Sensor Distribution and Adaptation
Navies have always relied on distributed communication because of the dis-
tances between vessels at sea. Line of sight (LOS) coded signaling by flag sema-
phores and nighttime signal lights preceded LOS radio communications be-
tween vessels and ship-to-shore communications. As ships conducted dispersed
operations in the second world war, the use of the electromagnetic spectrum
expanded, using VHF communications for two-way voice and coded messag-
ing between ships and to communicate with aircraft. By the 1960s, oscillator
technology enabled widespread move to longer range high frequency (HF) and
ultrahigh frequencies (UHF) for fleet communications. Satellite communica-
tions and navigation systems then enhanced the ability to communicate and
navigate worldwide, giving fleets significant gains in the coordination of battle
groups. The adoption of narrow beam extremely high frequency (EHF) com-
munications even allowed wideband communication from satellite to subma-
rines with EHF antennas mounted atop periscopes. The full use of the spectrum
for communication and fleet-wide access to satellite communications advanced
the ability to network and coordinate distributed vessels. TDLs and submarine
acoustic communications extended coordination with antiair and antisubma-
rine warfare platforms. The recent adoption of digital networking via TCP/IP
protocols has enabled standardization for adaptive, self-synchronizing, and ad
hoc networks—enabling global ashore-afloat communication of secure voice,
data, and control of unmanned platforms.
To distribute ISR and perform the fusion of sensors we must distinguish
several key technical elements of the distribution processing as distributed sen-
sor network (DSN):

131
132 Naval ISR Fusion Principles, Operations, and Technologies

• Distributed detection in a DSN is the process of integrating predetection

from multiple sensors to declare a detection of an object based on com-
bined signals. This we defined as level 0 fusion in Chapter 6; in this case
the sensors may be distributed and focused on a common faint source
of observable energy.
• Distributed tracking in a DSN consists of a set of processing nodes col-
lecting data from sensors; when the nodes communicate, each node fus-
es the information received from other nodes with the local information
to update its estimate on the state of the multiple target objects being
tracked [1].
• Distributed tip-cueing or handoff refers to the process where one sensor
in a DSN communicates a detection or track to another sensor to cause
it to acquire, detect, identify, or continue (handoff ) tracking the target
object.
• Distributed identification is the process of obtaining sensor information
from multiple sensor types (either concurrently or sequentially in time)
to identify a target object by the combined signatures to discriminate the
identity of a target object.

In this chapter, we develop the role of digital networking to enable ful-

ly distributed operations to achieve the vision of network centric fleets with
distributed ISR fusion. We illustrate with the background to the U.S. Navy
development and implementation of network-centric operations. This vision
was promoted in the late 1990s as a conceptual approach to moving from in-
dividual platform-centric warfighting (e.g., a vessel and its organic sensors, C2,
and weapons) to a network-centric approach where platforms, sensors, a data
network, weapons, and C2 are distributed to perform as a distributed system.
The NCW concept was defined as:

An operational concept for conducting warfare that 1) shifts the focus

of control from the weapons platform to the information network that
connects platforms, 2) shifts focus from independent warfighting actors
to a continuously adapting ecosystem of warfighters, and 3) operates on a
strategy to adapt and survive using a dynamically changing ecosystem [2].

The NCW concept, championed in the United States by Vice Admiral

Arthur Cebrowski would leverage advanced information technology to improve
sea power efficiency and capabilities. By 2001 the Navy developed a report,
“Network Centric Operations (NCO), A Capstone Concept for Naval Op-
erations in the Information Age” that articulated the naval benefits for flexible
tactics at sea and ashore [3]. In this timeframe, Cebrowski was promoting a new
 Sensor Distribution and Adaptation 133

way of war that included dispersed forces—noncontiguous operations that pre-

figured DMO [4]. The DoD NCW concept development period (1997–2003)
included the development of a value chain to explain how networked opera-
tions contributed to combat capabilities. The value chain analysis included case
studies, simulations, and real-world experiments to validate the concepts. Data
standards and metrics were developed to measure the benefit of future NCW
implementations [5]. The U.S. DoD promoted research in NCW through its
Command and Control Research Program (CCRP) publications and confer-
ences [6].
By 2002, the Navy established the implementation approach called
FORCEnet, “the operational construct and architectural framework for Naval
Warfare in the Information Age which integrates warriors, sensors, networks,
command and control, platforms and weapons into a networked, distributed
combat system, scalable across the spectrum of conflict from seabed to space
and sea to land” [7]. FORCEnet initiated platform and equipment roadmaps
that were driven by operational analyses, modeling, simulation, and war gam-
ing, followed by naval and joint force experiments. Incremental technology
improvements were made with sensors, data links, satellite sensing and com-
munications, data fusion processing, and visualization of a common operating
picture.
By 2020 the Navy committed to a Naval Operational Architecture
(NOA) including new networks, data standards, and formats, with battle man-
agement tools to be fielded and evaluated on a carrier strike group. NOA in-
cludes infrastructure based on cloud and edge computing, augmentation with
AI capabilities, platform extension with unmanned vehicles and vessels, and full
seabed-to-space interoperability. The scope and scale of an ambitious NCW
implementation across a fleet the size of the U.S. Navy has moved from con-
cept to implementation over the span of two decades and is projecting another
decade to reach completion.
In the next sections we extend the data fusion concepts of Chapter 6 to
perform network sensing and control in support of NCW.

7.1 Sensor Networks and Grids

Prior to DMO, the fleet operated primarily in clusters of vessels or battle groups
with an aircraft carrier encircled by cruisers and guided missile destroyers pro-
viding air defense; destroyers with towed array sonar and submarines provided
ASW defense. Sensors were local (organic sensors to the fleet), and coordination
of sensors was by relatively short range, even LOS links. The introduction of
DMO distributes vessels and their sensors far apart beyond line of sight (BLOS)
while adding remote unmanned sensors in space, the air, and on the surface.
134 Naval ISR Fusion Principles, Operations, and Technologies

The architecture of the networks (Table 7.1) that interconnect the sensors and
their users refers to the topology of the links between communication nodes:

• Links refer to communication channels that pass data between nodes;

• Nodes refer to the communication sources or sinks of two types:
• Communication endpoints nodes are sources that transmit data sen-
sors, (human voice, digital data sources, C2 data, etc.) or sinks that
receive data (processing, display, or other users of data);
• Redistribution nodes provide relay of data from one link to another,
to enable networking of links from sources to sinks.

Networks of links for digital data enable the distribution of command,

control, sensor data, Blue and Red force location data, common operating pic-
tures, and other mission data between source platforms (e.g., seabed-to-space
sensor platforms) and C2 platform to conduct operations.

Table 7.1
Link and Network Topology Structures
Structure Description of Use in Data Fusion Applications Example Naval Uses
Point-to-point The most basic link between 2 nodes (e.g., a sensor Sensor node to a sensor
and fusion processor, or between 2 data fusion processor.
processors).

Star Peripheral sensor nodes are connected to a Multiple sensors on a

central fusion processing node (hub). The hub may common platform with
rebroadcast data to all peripheral nodes on the a local fusion processor.
network. There is no direct traffic between nodes. This is centralized data
fusion.
Ring Every node has exactly two links connecting it to Sensor array with
additional nodes with each node acting as a repeater limited communication
to relay the data to the adjacent node. Each node operating in a line.
can be a sensor node, or a sensor + fusion node. The
failure of one node can impact the entire network.
Tree, Individual peripheral sensor nodes transmit to and Integration of many
hierarchy receive from one other node only, toward a central multiple sensor
node, and are not required to act as repeaters or platforms.
regenerators. This can be a hierarchical connection
of star subnetworks that send data to a central data
fusion node.
Mesh A network with at least 2 nodes with 2 or more paths A fully or nearly fully
between them. Every node has a dedicated point-to- connected set of
point link to every other device. sensors and data fusion
processors.
 Sensor Distribution and Adaptation 135

Naval communications have relied on a large variety of voice and data

links to transfer data between aircraft and vessels afloat, and to ashore com-
mand and intelligence centers. This heterogeneous collection of links formed
hybrid topologies that required data translation between diverse protocols to
move sensor data. Navies have moved to digital networked protocols (TCP/
IP) seeking near-mesh architectures to enable a more complete exchange of
sensor data (as well as common operating pictures and command data) across
the diversity of links (Table 7.2). These links require assured communications
in contested environments, interoperability with coalition forces, and security
(confidentiality, integrity, and availability). The myriad of links from seabed to
space for a deployed fleet require a significant adaptive network management
capability.
The organization of these many links can be identified in the following
categories:

• Afloat-to-ashore networks: Satellite links enable distant fleets to commu-

nicate with ashore modes (command centers, maritime intelligence cen-
ters, data storge, or other services).
• Tactical wide area networks (WAN) provide surface, ship, submarine, air-
borne, tactical-shore, and gateway services to shore-based WANs. These
nets use the higher bandwidth satellite links, VHF/HF links, and TDLs
to collect sensor data to processors and users, and to distribute common
operating picture and command data.
• Mobile user terminal networks are narrowband networks to mobile user
terminals (e.g., marines, special forces, or ground forces) from handheld
to small unit terminals.
• Advanced wireless networks: In the next section we introduce advanced
generation wireless technologies that will enable finer-mesh networks in
the future.
• Undersea networks: Acoustic, fiber-optic, and short-range RF links be-
tween a sensor field involving a variety of fixed (e.g., seabed arrays) and/
or moving nodes (submarines, towed arrays, UUVs, etc.) are employed
to conduct surveillance, detection, and localization of submarines or
mines. These systems are often cued by an initial detection, and then
multiple sensors perform a wide area search to secure multiple detec-
tions, localization, and finally classification and tracking. The deploy-
ment of a local network of sensors (e.g., sonobuoy array, or seabed bar-
rier array) are the best approach to high probability detection-tracking
after localization [9].
136 Naval ISR Fusion Principles, Operations, and Technologies

Table 7.2
Major Categories of Naval Optical and RF Communication Links
Communication
Links Capability Comments
VHF/HF fleet ·· Secure voice (analog and digital) and ·· STANAG4691 jam resistant
communications data modes; communications;
·· LOS surface-to-surface and long range ·· LOS comm to manned,
surface-to-air; unmanned air and surface
·· Data rates up to 1 mbps. vehicles.
HF long range ·· Long range extended line of sight (ELOS) ·· Viable alternative to
capability to 4,000 km; SATCOM when outside the
·· 3-kHz narrowband and 48-kHz wideband footprint of SATCOM (e.g.,
channels with 4G HF waveforms; northern latitudes);
·· Data rates of up to 240 mbps. ·· Mil-Std-188-110C/D &
STANAG 4539 App H provide
IP networking.
TDL heterogeneous ·· NATO Link 16; ·· Time-division multiple
HF/UHF networks ·· Data rates up to 2 mbps; access (TDMA) tactical data,
·· NATO Link 22; C2, mission, and weapons;
·· Data rates up to 12.7 mbps in UHF ·· Provides theater-level
mode. connectivity.
Ground to subsurface ·· Low bandwidth one-way instructions to ·· VLF (3–30 kHz) penetrates
UHF/ELF RF data submarine to establish a different form water to a few tens of
links (one-way to of two-way communication link (<10 bps) meters, EHF (3–300 Hz) to
submarine) hundreds of meters.
Subsurface RF, optical ·· Short-range 30–100m optical links >1 ·· Underwater links between
and acoustic data mbps submarines, arrays, and
links ·· Fiber optical links up to 10 km, >2 gbps; UUVs.
·· Acoustic com (ACOMM) links are
not covert but can achieve ranges
of 30–40 mi.
TDL over P-LEO for ·· Provides beyond LOS (BLOS) ·· Satellite-based TDMA
BLOS data link experimental TDL; terminal [8] with global
·· Link 16 terminals on P-LEO constellation coverage.
to enable persistent overhead access.
Geostationary (GEO) ·· Multiband (X, Ka) channels on geo- ·· GEO satellites orbit at
SATCOM to very small to-surface beams; geo satellites have altitudes of 35,786 km
aperture terminal multiple beams to provide shaped with an orbital period of 24
(VSAT) coverage areas; hours; round-trip latency is
·· Round-trip latency <550 ms; approximately 477 ms;
·· Data rates up to 250 Mbps down/20 ·· MEO satellites orbit at
Mbps up. altitudes of 2,000–35,786
P-LEO and medium ·· Near global coverage Internet Protocol; km with orbital periods
MEO SATCOM ·· Low round-trip latency <50 ms; between 127 minutes to
·· Data rates of up to 100 Mbps down/3 24 hours; RT latency is
Mbps up. 27–477 ms;
·· LEO satellites orbit at
Optical LaserCom ·· Very narrow beams, with minimal altitudes of 160–2,000
links satellite-to- divergence reducing ability to intercept; km with orbital period of
satellite crosslinks, ·· More power can be delivered to the 88–127 min; RT latency is
satellite-to-airborne, receiver compared with RF; approximately 2–27 ms.
and ground ·· Very high data rates >5 Gbps.
 Sensor Distribution and Adaptation 137

Note that the two categories of satellite links (GEO and P-LEO SAT-
COM) provide alternative channels in which bandwidth, latency, capacity
(data rates), and footprint (coverage) must be traded off to properly apply each
application. In addition, both military (e.g., MILSATCOM) and commercial
(COMSATCOM) capabilities, may be appropriately applied to provide SAT-
COM diversity for operation in contested, degraded, and operationally limited
(CDO) environments [10].

7.2 Advanced Wireless Networks

While the prior section focused on traditional dedicated naval communication
networks (operating on military and commercial communication standards)
a recent area of intense interest is the adoption of advanced mobile wireless
communications to provide high bandwidth, low-latency, multimedia content
rich communications at sea. Mobile wireless communication technology has
evolved to the point where it is feasible to incorporate it into the NCW enter-
prise and realize the necessary technical requirements for data exchange.
The fifth generation (5G) of mobile telecommunication systems and the
future 6G is the follow-on standard that provides very high data rates with very
low delays (0.1 ms). The goal of these next generation standards is to allow hy-
perconnected operation that will enable connection between human users and
the Internet of Things (IoT). 5G offers greater data throughput and lower la-
tency compared to previous generations that include 2G (GSM), 3G (UMTS),
and 4G (LTE and LTE-A), but it is also closely aligned with the growth of the
IoT and affords technologies that support large-scale machine to machine com-
munication (Table 7.3).
In the development of the 5G standards, the 3rd Generation Partnership
Project (3GPP) and the 5G standards organization conducted a feasibility study
on new services and markets technology enablers called the SMARTER report;
the study identified a set of use cases that greatly informed the resultant techni-
cal specifications of 5G [11]. The predominant use cases identified were:

• Enhanced mobile broadband (EMBB) are applications and services that

are data-driven, requiring high data rates and low latency across a wide

Table 7.3
Key Characteristics of Wireless Technology Generations
Generation Download Data Rate; Link Latency Density (Links)
4G Download 1 Mbps; latency 60 ms to 98 ms 105 devices/km2
5G Download 10 Mbps; latency under 5 ms 106 devices/km2
6G Download 95 Gbps; latency 1 ms 107 devices/km2
138 Naval ISR Fusion Principles, Operations, and Technologies

coverage area. Examples include support for mobile broadband, UHD/

holograms, and virtual presence technologies.
• Mission critical/ultrareliable low-latency communications (URLLC) are
applications and services that have strict requirements on latency and
reliability for mission critical communications. Examples include in-
teractive games/sports, industrial controls, drone/robot/vehicle, remote
surgery, tactile internet, and emergency services.
• Massive machine-type communications (MMC) are applications and
services that support many devices in a small area and exchange data
infrequently. Examples include IoT devices, subway/stadium services,
eHealth, wearables, and inventory control.

From these use cases, a set of derivative requirements were postulated; the
general use cases can be compared to DMO maritime applications (Table 7.4).
In the DoD Digital Modernization Strategy, 5G is highlighted as a key
enabling technology, with several of the benefits identified, including the “abil-
ity to deliver fiber-like speeds to end-user devices, improved performance at
network cell edge, low-latency performance (<2ms radio latency), and greater
spectral efficiency” [12]. The strategy further identifies foundational technolo-
gies including: “improved beamforming via smart antennas, network densifica-
tion via small cells, native support for diverse spectrum types (licensed and un-
licensed), and increased network intelligence via software defined networking
(SDN)” [12]. In the DoD 5G Strategy Implementation Plan, the Department
also promotes a 5G-enabled concept of operations to leverage the “transforma-
tional speed and connectivity of 5G” [13] and early adoption of the capabilities
associated with the EMBB, URLLC, and MMC use cases.
From a maritime perspective, 5G offers ubiquitous coverage benefits to
the maritime transportation industry [14], and to the naval deployment and
adoption of autonomous vessels, air vehicles, and unattended sensor networks.
Unmanned or remotely piloted commercial vessels enable extended at-sea
availability, and thus greater transport volume than currently achievable with
traditional cargo ships. Inventory control and monitoring aboard such vessels
enabled by 5G’s MMC capabilities is another area where 5G benefit can be
realized. With IoT sensor-equipped containers, this dense onboard network of
assets, can be accurately located, tracked for chain of custody, and monitored
for health and wellness. Additionally, with IoT-connected sensors aboard, vessel
performance can be used for route and speed optimization and for predictive
maintenance.
These maritime benefits extend to DMO applications when ubiquitous
coverage integrates satellite constellation networks into 5G networks. This
requirement is addressed by 3GPP in its global satellite overlay use case that
 Sensor Distribution and Adaptation 139

Table 7.4
Summary of Operating Characteristics for 5G Use Cases
Naval Maritime
Use Case Data Rate, Latency Traffic Density Applications
EMBB: Very high data rate (e.g., High traffic density Real-time sensor and
Characterized by very up to 10 Gbps) when the (Tbps/km2) imagery data transfers
high traffic, high bitmobile user is moving High connection Air-air, air-satellite, and
rate with moderate slowly density of user air-ground real-time
reliability and Very low latency, reliable equipment: communications
accuracy low latency for high- 200–2,500 devices/
speed connections (e.g., km2
aircraft)
Ultrareliable Real-time ultrahigh High traffic density Critical targeting
low-latency reliability real-time with distribution (e.g., information transfers;
communications high availability for 10,000 sensor target locations, tracks,
(URLLC): mission critical traffic devices /10 km2) and targeting data
Characterized by low Low latency (delay) in updates in real time
latency, ultrahigh some dense cases 1–10
reliability, and ms and full round-
precise position trip latency less than
centimeter 100–150 ms
MMC: No low latency High traffic density Unattended sea surface
High positioning requirement, but efficient massive connections and other sensors
accuracy (e.g., 0.5m) resource and signaling to up to 1 million reporting data
support low power with connections per km2; Autonomous surface
limited communication appropriate for large and air vehicle
requirements numbers of locally environmental and
dense devices noncritical data
reporting

envisions a constellation of LEO satellites, which form a nonterrestrial net-

work, where each spacecraft is equipped with a base transceiver station (a next
generation NodeB, gNB) and is interconnected with other satellites in a con-
stellation via intersatellite links. This constellation of satellites contributes to
a single 5G system and provides 5G access to at-sea users. This augmentation
allows global 5G coverage and Figure 7.1 depicts user access to a data network
via a nonterrestrial 5G network enabled by the LEO satellite constellation. The
user equipment (UE) connects to the spaceborne base transceiver station (gNB)
supported by the next generation core (NGC) equipment to access a terrestrial
data network to ashore user equipment.

7.3 Functional Nodes on the Maritime Network

A RAND study of the U.S. Navy growing sensor data networking and analytic
needs estimated the exponential growth of sensor data as unmanned sensor
platforms grew and sensor data volume increased. The 2014 report warned:
140 Naval ISR Fusion Principles, Operations, and Technologies

Figure 7.1 Nonterrestrial network featuring an access network serving UEs and based on a
satellite with a spaceborne base transceiver station on board [15].

Unfortunately, much of the data collected by ISR platforms never reaches

the analysts who need to see it—in fact, as little as 5 percent may be getting
through. In the case of analysts working afloat on ships, a large part of
the problem is extremely slow download times caused by bandwidth
and connectivity limitations. Whereas an ashore analyst working with a
100-gigabits-per-second transfer speed can download a terabyte of data
in four minutes, an afloat analyst working with a 40-megabits-per-second
transfer speed needs three days to do so [16].

The study concluded that SOA-based or cloud-based computing architec-

tures were required with source data-tagging, wideband data links, and multi-
level security to manage the massive volume of maritime data.
In Section 7.2 we introduced the fundamental links and nodes for com-
munication networks; in this section we expand the discussion to include the
functional nodes at the application layer of the maritime ISRT network. At this
higher level, we distinguish the ISRT functional nodes (Table 7.5) that do not
just route communications but also sense phenomena, collect sources of in-
formation, process information (data fusion), and then act upon the maritime
environment [17].
We illustrate a very simple functional network (Figure 7.2) that includes
just a small set of action platforms to show the interactions between three of the
five categories of nodes.
Figure 7.2 shows the network in the following simple scenario:

• An unknown target ship is detected by the acoustic sensor S1 on a UUV

platform that relays the detection and signature to a surface buoy or
USV, R1 that relays the information to a LEO relay constellation. Relay
satellites R2… Rn… R3 pass the detection report to an afloat fusion
node F3 and to an ashore fusion node F4.
• An area patrol aircraft is vectored to within range (command link not
shown) to observe the target with synthetic aperture radar S2, electronic
support measures S3, and electrical optical imagery if feasible S4. A fu-
sion processor onboard the aircraft fuses the sensor data F1 and passes
 Sensor Distribution and Adaptation 141

Table 7.5
ISRT Major Functional Nodes
Node Description Naval Example
Sensor or source A node that produces sensor data Radar, IFF, sonar, EW, sensors;
or source (e.g., HUMINT, or OSINT) Imagery, SAR, SIGINT sensors;
information for distribution HUMINT sources, force trackers.
Relay Node that only relays information from Satellite relay (GEO, LEO);
node-to-node across a communication Tactical data link TDMA protocol;
network; information is not processed Any TCP/IP router in a WAN.
Processing-fusion A node that receives sensor and Sensor processing and data fusion
source information from multiple nodes on any platform;
nodes to process (detect, track, ID) All-source analyst units data fusion
and fuse (combine, control sensor/ nodes that combine multiple INT
sources), then forwards results to sources;
other nodes MultiINT processors.
C2 A consumer node that accepts C2 processors, common operating
information to support C2 functions picture display processors;
C2 units that direct ops and weapons.
Weapon A consumer node that receives Weapon system (e.g., missiles,
information to influence the operating guns, mines) and supporting weapon
environment platforms.

Figure 7.2 A simple functional network for a small action unit.

track and ship class data are through the LEO relay constellation to the
afloat command ship and ashore fusion processors F3 and F4.
• LEO remote sensing satellites with SAR S5, and EO S6, sensors are
directed to observe and collect on the target (command link not shown)
142 Naval ISR Fusion Principles, Operations, and Technologies

and report imagery through the LEO relay constellation to the afloat
command F3 ship and the forward destroyer F2.
• The ashore and afloat ships F2, F3, and F4, coordinate the common
operating picture and fusion processes through the geospatial relay sat-
ellite R4. The issue fusion processor integrates information from other
maritime sources in fusion processor F5. The result is that the command
ship and the forward destroyer have the combined information about
the target.

This simple example of a mesh network has 18 links and 16 nodes, and
only describes the most minimal set of sensor-reporting transactions that would
occur in such an activity; in an actual scenario repeated and continuous sensor
observations to maintain custody of the target would significantly increase the
transactions. Sensor control, platform maneuver guidance, C2, and weapon
command transactions have also not been included in this simple example.
The mesh network of even a moderate DMO naval unit can become
complicated, and its adaptive transactions can become complex. Consider the
network example (Figure 7.3) for a moderate surface action group. The distrib-
uted force covers an area of 500,000 mi2 with a surface to-seafloor blue-water
depth 5,000 ft with littorals (shallow water areas) [18]. The DMO Force is
distributed over 62,500 sq nautical miles (250 × 250 nm area) and includes a
carrier strike group comprised of:

• 1 carrier CV; 8 destroyers DDG;

• 2 attack submarines SSN;
• 2 maritime patrol aircraft P, and 2 air surveillance aircraft AEW;
• 4 forward-deployed UAVs;
• 4 forward-deployed UUVs;
• 4 forward-deployed fighter/attack aircraft (VF).

Figure 7.3 distinguishes the network nodes with the carrier (solid node
at center) encircled by widely distributed destroyers (double circle nodes) and
the forward deployed air assets. The example net includes 157 links but notably
does not include numerous other nodes and links: USVs, undersea links be-
tween undersea vehicles, multiple RF links available to undersea vessels, buoys
and deployed surface sensors, and seabed sensors. The net also simplifies the
dynamic persistent LEO relay constellation as a single node, when, in fact, it is
a continually changing set of nodes and relay nodes. It is important to note that
this included only the functional communication links, but did not distinguish
 Sensor Distribution and Adaptation 143

Figure 7.3 Composition and network for SAG.

sensor, fusion process, and relay nodes as in Figure 7.2. The figure illustrates the
complexity of the most basic set of nodes for a DMO strike force.
The network figures (Figures 7.2 and 7.3) also illustrate the challenge
of designing the structure by which multiple sensor nodes and fusion nodes
relay and combine information. This consideration is important for maritime
networks and well as wireless and mobile ad hoc networks. We introduce this
distribution of sensors and fusion processes in Section 7.4.
144 Naval ISR Fusion Principles, Operations, and Technologies

7.4 Distributed Data Fusion in Network Operations

Because sensors are distributed across a large network and many platforms will
maintain their own fusion processing of their local sensors (a patrol aircraft,
a UAV, a ship with multiple sensors, etc.) it is necessary to be able to per-
form data fusion and sensor management in a decentralized or distributed data
fusion (DDF) system [19]. We also distinguish homogeneous sensor networks
with common sensors (e.g., acoustic arrays or radar networks) and heterogenous
networks which include different sensors that contribute to data fusion with
multiple features to aid in detection, identification, and counter deception. Of
course, a centralized data fusion system would have a star network structure
with all sensors contributing to a central data fusion process; this can be per-
formed on the local platforms mentioned previously, but it is not feasible for
the DMO force.
Distributing the fusion process across diverse platforms and heteroge-
neous sensors is a significant challenge due to several factors, including:

• Alignment of out-of-sequence observations: Independent sensors may ob-

serve a common target at different times and report the sample across
different network paths, resulting in the observations being received at
the fusion node out of sequence. The distributed processor must place
the samples in proper time sequence to correctly associate the samples
and estimate the state of a target.
• Distributed estimation errors: The estimation process in a fusion node
combines the information from sensors, or from another fusion proces-
sor to produce refined estimates of several kinds:
• Estimated static location;
• Estimated identification by a classification function;
• Estimated dynamic track (a model of kinematic behavior, such as a
Kalman filter model);
• The estimate can be contaminated by unaccounted for correlations
(dependence) between measurements, correlated estimation errors
due to process noise (target behavior) or measurement noise (sensor),
or errors in the association of measurements.
• Data recycling: Once sensor data is correlated and combined in a fu-
sion processor (FP1, such as on a patrol aircraft) and is passed to an-
other fusion processor FP2 (e.g., on a destroyer) there is a risk of the
information being recycled (recorrelated and recombined) if any of the
contributing sensor data are also provided to the subsequent fusion pro-
cess (double-counting) and incorrectly assumed to be independent, this
will introduce a redundancy and present a cross-correlation error that
 Sensor Distribution and Adaptation 145

will contaminate subsequent combination processes (and exaggerate the

confidence in results).
• Distributed sensor coordination and control: To effectively apply the net-
worked sensors it is desired to orchestrate the sensors to assure coverage
across multiple targets by scheduling M sensors to observe N targets and
specified intervals. This process of sampling control across a network
introduces sensor control and data return delays that further complicate
the desire for synchronized target sampling, propagation delay reduc-
tion, and optimal coverage of the N targets.
• Imperfect association: The integrity of a track depends on all measure-
ments in the track originating with the same tracked object. If the
sensor-based tracking algorithm incorrectly associates measurements of
one entity with the track of another entity, this introduces process noise
and corrupts the estimate. When the error is recognized, the wrongly
assigned measurements must be removed and the tracks corrected ap-
propriately.

These factors require the designers of distributed ISR systems to consider:

(1) the structure of the distributed process and allocation of fusion nodes (see
examples in Figure 7.2), (2) careful metadata tagging of sensor and fusion data,
(3) understanding of the statistical noise properties of the links, target processes,
and sensor measurements, and (4) link latencies (end-end delays). Whereas Fig-
ure 7.2 shows the structure of sensor and fusion nodes, Figure 7.3 shows the
available data links (with different capacities, data rates, latencies, and other
properties that must be considered). The two are related since information must
flow over physical and logical communication links. These links are designed
to support a particular concept of operations for which they are optimally de-
signed. However, multiple ways exist to repurpose links to facilitate the flow of
information. Within the constraints imposed by the range and bandwidth of
the links, they may be able to support multiple information flow architectures,
thus they inform, but do not dictate, options for fusion strategy. We have enu-
merated the basic communication links-to-fusion node structures in Figure 7.4.
In the example in Figure 7.2, the shore node and multisensor platforms
(patrol, UAV, UUV, VF, etc.) may have central fusion node structure on the
platform. Some elements of the force may have a hierarchical structure (e.g., the
unmanned platforms to their command vessels) and the complex interaction
between large vessels and ashore may be a distributed/mesh structure.
Centralized fusion strategies involve funneling all sensor measurements
to a central decision node. This node performs all association of measurements
with tracks. The advantages of this architecture are that there is no data recy-
cling—all measurements are processed once, and because of this it offers the
146 Naval ISR Fusion Principles, Operations, and Technologies

Figure 7.4 Example networked data fusion structures [20].

best theoretical performance regarding minimizing errors. There are many dis-
advantages though. Obviously, the single fusion node represents a single point
of failure. In addition, the convergence of all sensor feeds to the central node
creates a communications bottleneck, which limits scalability to the bandwidth
and computational capacity of the central node and network links.
Decentralized/distributed fusion is the opposite end of the spectrum; ev-
ery node in a mesh can share its sensor data directly with every other node.
Each node then performs its own fusion based on its own local sensor mea-
surements and the measurements of all other nodes. This obviously maximizes
redundancy, which increases resilience, but redundancy has other costs. In this
architecture, the tracks formulated at each node may contain redundant infor-
mation as nodes may have access to common sensors. The most salient informa-
tion influences track formation and as those tracks are shared, not as additional
measurements, but as the conclusions of the fusion process at EACH node they
may result in double-counting—using a sensor measurement more than once as
if the two observations were independent. Furthermore, the ideal of a perfectly
 Sensor Distribution and Adaptation 147

interconnected mesh of nodes is rarely, if ever, realized. In practice, links will

be broken, information will arrive out sequence, and the nodes will have an
inconsistent picture of the situation.
In most realistic naval scenarios, a hybrid approach will be required. Naval
engagements are, in general, geographically dispersed. The geography will cause
the available sensors and platforms to be within range of some fusion nodes and
out of range of others. The most reasonable architecture for such situations is
likely to involve local, centralized fusion where it is optimal and distributed fu-
sion of the local fusion centers where that is optimal. Because the aim is to even-
tually arrive at one best interpretation at the top level of command (in our case
the shore node), this architecture approach is usually referred to as hierarchical.
A method for choosing which architecture to employ has been developed
that uses ten selection criteria (below) to be maximized or minimized [21]:

1. Assessment accuracy (MAX);

2. Complexity in correcting misalignment of measurements (MIN);
3. Latency (MIN);
4. Robustness/survivability (MAX);
5. Computational resources required (MIN);
6. Communication bandwidth required (MIN);
7. Implementation complexity (MIN);
8. Size, weight, and power (MIN);
9. Cost (MIN);
10.   Energy cost (MIN).

With these ten objectives in mind, any architecture can be considered.

Each architecture can further be parameterized in terms of number of hierarchy
levels, number of processors per hierarchy level, and architecture type at each
hierarchy level. Thus, we can describe the architecture selection decision as a
multiobjective optimization problem based on the aforementioned factors.
Selecting the right architecture is the first part of designing a system for
distributed sensing. Equally important is the selection of the proper fusion al-
gorithms. The goal of MultiINT fusion is to fuse the combined measurements
from multiple sensors over time into a track that ideally describes with accuracy
the position, velocity, entity type, and entity state. Before we discuss algorithms,
we must first define our terms:

• State refers to a possible configuration of an abstract state space (the set

of all possible configurations of a system). In a practical physical system,
the static kinematic state is defined as the position and orientation of an
148 Naval ISR Fusion Principles, Operations, and Technologies

object (e.g., a ship) in physical space while the dynamic state includes
the associated velocities. Commonly, the kinematic state of a ship or
aircraft, for example, includes the location in a coordinate system (e.g.,
x, y, z) and the velocity vector.
• Measurement is an observation recorded by a particular sensor at a par-
ticular time. Metadata for each observation should be available; particu-
larly important is the time at which the observation was made as we shall
see later in this section, in Figure 7.6.
• Track is a history of measurements associated with a single entity of in-
terest and stored in a time ordered sequence. Tracks can be maintained
at the sensor level or the system level.
• Sensor track is a track maintained by a single modality sensor, even
though multiple sensors may exist on a single platform like a UAV,
the sensor track would refer to just one of those sensors. The sensor
algorithms determine how to associate the individual measurements
into a track.
• System track is a track of a single entity that is formed at a fusion node
by fusing data from more than one sensor and/or other fusion nodes.
• Tracklet is a short subset of measurements representing a finite time
series segment of an ongoing track.
• Process noise is noise that is inherent in the behavior of the observed en-
tity. This could be due to environmental clutter or other nearby entities.
An example would be incorrectly including a measurement on entity B
in a track associated with entity A, corrupting the estimate of the state
of A.
• Covariance matrix contains elements quantifying the measurement un-
certainty associated with a measurement due to process noise and obser-
vation noise.

The most common statistical tracking models are based on some form
of the Kalman filter [22]. The basic Kalman filter (Figure 7.5) seeks to make a
prediction of a state at time k, xˆk based on a series of measurements, or at least
the most recent measurement at time k–1, zk–1. The filter assumes that the true
state can be modeled by a state transition matrix, F operating on the current
state, plus some process noise, wk:

x k = Fk x k −1 + w k (7.1)
 Sensor Distribution and Adaptation 149

Figure 7.5 The basic Kalman filter functions.

Of course, the actual state is not known to the ISR system which is pas-
sively monitoring an uncooperative target. All that can be known is the observa-
tions that it makes, where the observation is given as:

z k = H k z k −1 + v k (7.2)

where H is the observation model and v is the observation noise. The role of the
observation model is to map observations to derived states. From an observa-
tion, or more often from a series of observations, the Kalman filter computes
the predicted state, and the estimated covariance matrix, P as follows:

xˆk |k −1 = Fk xˆk −1|k −1 (7.3)

Pk |k −1 = Fk xPk −1|k −1Fk−1 +Q k (7.4)

where Q is the covariance of the process noise. These predictions can be used to
keep targets in track for weapons engagement or could inform the ISR system
where to look next if resources are scarce. If a track prediction seems relatively
nonvolatile, less resources need to be used to track it than a target which ex-
hibits highly dynamic behavior with large, forecasted covariance. Later in this
section, we will explore how variations on the Kalman filter are used in DDF
and will consider a few other algorithms as well.
Before discussing algorithms, we need to consider where and at what level
the different kinds of information can be fused. A single sensor only has access
to its own measurements, thus the only option it has is to fuse the measurements
it collects over time into a set of tracks. A single sensor platform with multiple
sensors has more options. It is in theory possible to fuse the measurements of
each sensor together and then form a track from the fused measurements. In
150 Naval ISR Fusion Principles, Operations, and Technologies

practice, however, sensors may collect data on different time scales and in many
different formats, so it is much more common to let each sensor to form its
own tracks and then fuse those tracks. Sensor-to-sensor direct fusion is pos-
sible, however, in this example, the platform functions as a central fusion node
producing system tracks by sensor-to-sensor fusion.
A variety of combinations of system-to-system and sensor-to-system fu-
sion become possible as systems get larger. For instance, consider the shore node
(F4) in Figure 7.2 which receives system tracks from F2, F3, and F5 and sensor
tracks from satellites S5 and S6.
Finally, we need to understand the concept of information flow and how
that affects fusion algorithms. Information flow diagrams, like that depicted
in Figure 7.6 show a series of measurements from two sensor systems, each
maintaining a track on the same target, which are periodically fused into a
system track. Obviously to construct an information flow, measurements must
be source-labeled and time-tagged (pedigree, as introduced in Chapter 6) to
be processed in correct order. Information flow is used to decide how many
measurements to collect into a tracklet before fusing with other tracks. It can
also be used to avoid the data double-counting (or recycling) issue mentioned
earlier in this section. This diagram (Figure 7.6) illustrates how dependencies
arise when measurements are used to generate a track estimate and that (system)
track estimate is used to fuse with another (sensor) track containing the same
measurements, resulting in overconfidence placed on those measurements.
Notice in Figure 7.6 that both sensors experience dropouts in coverage,
but the system track maintains a combined track estimate in the presence of
intermittent data. This example could represent two radars, one on a ship and
one on a distant aircraft tracking an aircraft or ship but experiencing dropout
periods; the system track provides sustained custody of the target in this simple
example.
We are now able to discuss some of the more important fusion algo-
rithms applied in DDF and the situations in which they are applicable. The
simplest fusion algorithm we may consider is a centralized Kalman filter, where

Figure 7.6 Sequence of sensor measurements, sensor tracks, and a combined system track
for two sensors (figure based on [23]).
 Sensor Distribution and Adaptation 151

all observations are sent to a single fusion node. We can treat these observa-
tions as a single observation vector and the fusion equation then reduces to the
simple Kalman filter in (7.3) and (7.4). In the more realistic case of a hybrid
or hierarchical architecture, a decentralized version of the Kalman filter needs
to be employed. Each fusion node would employ a centralized Kalman filter
at the local level. If we can assume that the tracks do not interact, then the
overall state of the situation can be represented by the naïve combination of
the tracks produced by the local filters. However, if more than one sensor or
fusion node is observing the same track, then the shared information must be
communicated and reconciled. Consensus based versions of the Kalman filter
can be used in this case, with variations building consensus on measurements
or on the magnitude of shared information [24]. A simple way of reconciling
tracks is to compute a linear combination of the reported tracks. The weighting
of the linear components can be as simple as weighted least squares fit if the
track observation errors are independent or could be complicated if shared error
sources are present, in which case the cross covariance must be computed [23].
Earlier in this section we mentioned the problem of data recycling, which
can lead to overconfidence in estimates because some observations are effec-
tively counted multiple times. One way to deal with this problem is to re-
construct the data input and remove the double counted data. This is easy to
visualize if one considers the information flow diagram in Figure 7.2. Another
approach would be to restart filters after information is fused. For instance, a
sensor would compute a track until m measurements are received and the track-
let is fused with the system track. The sensor tracker then begins anew and com-
putes another tracklet based only on the measurements m + 1 through m + n,
and that tracklet—containing all ‘fresh’ information—is fused with the system
track. Conversely, the sensor tracks could continue to build upon all available
measurements and the system track could be reset at each fusion event [23].
We caution that these considerations assume a high, continuous sampling
rate on the target (relative to target velocity)—for example, multiple radars
observing a common target, or multiple satellites revising the target area very
often. But this is not always the case, particularly in DMO scenarios where
ships, aircraft, and their radars are widely dispersed, and the target is OTH.
DMO data fusion designers must consider the combined effects of sample rate
and latency (Figure 7.7). Delayed tracking (measured in seconds to minutes)
may be acceptable, but the combination of slow or irregular sample and high
latency may render data fusion infeasible.
We can see that there are many architectures and algorithms to support a
variety of fusions schemes. Selecting the appropriate tools will depend on the
topology of the problem—the distribution of sensors and entities in time and
space, the overlap between sensors, the connectivity and latency in the commu-
152 Naval ISR Fusion Principles, Operations, and Technologies

Figure 7.7 Regions of tracking feasibility and performance.

nications infrastructure, and the ability to properly account for and correct data
recycling and out of sequence data.
Two implementations of DDF for dense maritime applications (both
dense harbor monitoring with multiple sensor feeds) have been published with
explanation of the practical design considerations employed [25]. Both papers
explain the alternative architectures described previously and the considerations
in selecting an approach for their specific problem; one paper describes the ap-
proach to simulating alternative approaches and provides a comparison of the
results.

Endnotes
[1] Chong, C-Y, et al., “Distributed Tracking in Distributed Sensor Networks,” American
Control Conference, 1986, 1863-1868.
[2] Adapted from Cebrowski, A. K., and J. J. Garstka, “Network-Centric Warfare: Its Origins
and Future,” U.S. Naval Institute Proceedings, January 1998, pp. 28–35.
[3] Network Centric Warfare, Department of Defense Report to Congress, July 27, 2001.
[4] Cebrowski, A. K., “Defense Transformation and Network Centric Operations,” NDIA
Network Centric Operations Conference, April 2003.
[5] Though the NCW concept seems intuitive and was backed by analyses, it was not without
criticism. See, for example: Hammes, T. X., “War Isn’t a Rational Business,” Proceedings
of the U.S. Naval Institute, July 1998. Borgu, A., “The Challenges and Limitations of
‘Network Centric Warfare’: The initial views of an NCW sceptic,” Australian Strategic
Policy Institute (ASPI), 17 September 17, 2003. Giffin, R., and D. Reid, “A Woven Web
of Guesses, Canto One: Network Centric Warfare and the Myth of the New Economy,”
Proc. 8th International Command and Control Research and Technology Symposium, 2003.
Giffin, R. and D. Reid, “A Woven Web of Guesses, Canto Two: Network Centric Warfare
 Sensor Distribution and Adaptation 153

and the Myth of Inductivism,” Proc. 8th International Command and Control Research and
Technology Symposium, 2003.
[6] See key publications, such as Alberts, D. S., J. J. Garstka, and F. P. Stein, Network Centric
Warfare: Developing and Leveraging Information Superiority, Command and Control in the
Information Age, DoD Command and Control Research Program, CCRP Publication
Series, 1999. Alberts, S., and R. E. Hayes, Power to the Edge: Command and Control in
the Information Age, DoD Command and Control Research Program, CCRP Publication
Series, 2003. Moffat, J., Complexity Theory and Network Centric Warfare, DoD Command
and Control Research Program, CCRP Publication Series, 2003.
[7] Definition from July 22, 2002 CNO Briefing.
[8] Erwin, S., Viasat to Begin Integration of Long-Delayed Link 16 Military Com-
munications Satellite, Space News, April 21, 2022, https://spacenews.com/
viasat-to-begin-integration-of-long-delayed-link-16-military-communications-satellite/.
[9] National Research Council of the National Academies, Distributed Remote Sensing for Na-
val Undersea Warfare: Abbreviated Version, Washington, D.C.: The National Academies
Press, 2007, https://doi.org/10.17226/11927.
[10] “United States Space Force Vision for Satellite Communications (SATCOM),” United
States Space Command (USSPACECOM), January 23, 2020. See also “Proliferated-Low
Earth Orbit (P-LEO) COMSATCOM Services Request for Information (RFI),” United
States Space Force (USSF) Commercial Satellite Communications Office (CSCO), March
2021.
[11] “3rd Generation Partnership Project; Technical Specification Group Services and System
Aspects; Feasibility Study on New Services and Markets Technology Enablers; Stage 1
(Release 14),” European Telecommunications Standards, Institute 2017.
[12] Department of Defense Digital Modernization Strategy, July 12, 2019.
[13] Department of Defense 5G Strategy Implementation Plan, 2020.
[14] Harris, S., “Ship Ahoy: All at Sea with 5G,” Orange Business Services, February 05, 2021,
https://www.orange-business.com/en/blogs/ship-ahoy-all-sea-5g.
[15] 3rd Generation Partnership Project Specification 3GPP TR 38.811 V15.4.0 (2020-09).
[16] Porche, I. R., III, et al., “Data Flood: Helping the Navy Address the Rising Tide of Sensor
Information,” Report RR315, Santa Monica, CA: RAND, 2014, p. 14.
[17] These designations are based on Cares, J. R., et al., “Fundamentals of Distributed,
Networked Military Forces and the Engineering of Distributed Systems,” Naval Undersea
Warfare Center, NUWC-NPT Technical Report 11366, May 9, 2002. See also the earlier
work: Cares, J., Distributed Networked Operations: The Foundations of Network Centric
Warfare, Newport, RI: Alidade Press, 2021.
[18] For reference, the South China Sea is a marginal sea that is part of the Pacific Ocean,
encompassing an area from the Karimata and Malacca straits to the Strait of Taiwan of
around 3,500,000 km2 (1,400,000 mi2). The China Sea Basin, has a maximum depth of
5,016m.
154 Naval ISR Fusion Principles, Operations, and Technologies

[19] For a complete treatment of the applications for NCW, see: Distributed Data Fusion for
Network-Centric Operations, Hall, D., C-Y Chong, J. Llinas, and M. Liggins (eds), Boca
Raton: CRC Press, 2012.
[20] Based on examples in Chong, C. Y., “Distributed Architectures for Data Fusion,” Proc. 1st
Int. Conf. on Multisource-Multisensor Information Fusion, Las Vegas, NV, 1998.
[21] Jevti, M., N. Zogovi, and S. Graovac, “Multi-Sensor Data Fusion Architectures Revisited,
Information Society of Serbia,” 2019.
[22] For an in-depth treatment of tracking, see: Bar-Shalom, Y., Multitarget multisensor tracking:
Applications and Advances, Vol. I and II, Academic Press, 1990, 1992. Bar-Shalom, Y.,
L. Campo, “The Effect of the Common Process Noise on the Two-Sensor Fused Track
Covariance,” IEEE Trans. Aerospace and Electronic Systems, Vol. 22, pp. 803–805, 1986.
Bar-Shalom Y., and D. Blair, Multitarget Multisensor Tracking: Applications and Advances,
Vol. 3, Norwood, MA: Artech House, 2000.
[23] Chong, C-Y, S. Mori, K. C. Chang, and W. H. Barker, “Architectures and Algorithms for
Track Association and Fusion,” IEEE System Magazine, January 2000.
[24] Chong, C-Y, “Forty Years of Distributed Estimation: A Review of Noteworthy
Developments,” Proc. 11th Symposium Sensor Data Fusion, Bonn, Germany, 2017.
[25] See Ng, J., “A Distributed Sensor Network Architecture for Defense Against the Ship as
a Weapon in the Maritime Domain,” Monterey, CA: Naval Postgraduate School, 2011,
http://hdl.handle.net/10945/5698. Juliera, S. J., and R. Mittub, “Distributed Data Fusion
and Maritime Domain Awareness for Harbor Protection,” Naval Research Laboratory,
October 2006.
8
The Role of AI, Automation, and
Autonomy
While the applications of AI, automation, and autonomy are broad, in this
chapter we focus on the roles of these capabilities in ISR and its companion
process, C2 [1]. In addition, we must recognize that many automated ISR fu-
sion functions (especially correlation, association, combination, inference, and
estimation) have been applied in naval systems for decades and are now consid-
ered elements of AI.
For purposes of this chapter, we distinguish several key terms that refer to
a machine, a process, or a system [2]:

AI is a broad term that is most briefly described as the capability of a ma-

chine to imitate intelligent human behavior [3]. Human learning, reason-
ing, decision-making, adapting, planning, scheduling, and acting are the
kinds of behaviors that are the focus of AI applications that are of greatest
importance to the ISR enterprise. Note that we include the role of ML as
an important, but not separate element of AI.
Automation refers to some degree of autonomy, but control may be as-
sumed by human intervention. Such a system emphasizes efficiency, pro-
ductivity, quality, and reliability, operating without direct control, often
in structured environments over extended periods, and on the explicit
structuring of such environments.
Autonomy refers to the capability of a self-governing system and is condi-
tioned by the degree of self-governance. Full autonomy refers to a system
that is fully independent and operating without external human interven-

155
156 Naval ISR Fusion Principles, Operations, and Technologies

tion or influence. The system senses and responds to the environment,

adapting to changes to perform an objective capability. The DoD similar-
ly has defined autonomy as the ability of an intelligent system to indepen-
dently compose and select among different courses of action to accom-
plish goals based on its knowledge and understanding of the world, itself,
and the situation [4]. Machines, processes, or systems that exhibit com-
plete autonomy are referred to as being autonomous; those who exhibit a
degree of, but not complete self-governance, are called semiautonomous.
Unmanned surface vessels, underwater vehicles, air vehicles, and satellites
are all candidates for semi- and full autonomy to carry out well-defined
ISR missions. AI-enabled autonomy for these platforms is an important
contributor to DMO operations as they extend ISR reach, adapt rapidly,
and add increase the spatial distribution of ISR collection network.

We consider the following specific capabilities that are anticipated to

accrue to ISR and C2 operational effectiveness from AI, automation, and
autonomy:

• Enable expanded ISR sensing and sourcing: Autonomous sensor platforms

from seabed-to-space can efficiently collect and extract information, ma-
neuver to enhance collection and evade detection, and communicate
reports to the network.
• More effective uses of the highly diverse, large volume, and rich variety of
ISR sources and related contextual data include automated screening,
sensemaking, fusion, pattern discovery, and learning from high-volume
wide-variety MulitINT data for detection, discovery, recognition, and
targeting. While these capabilities have varying degrees of maturity, AI
capabilities with machine-learned models have significantly improved
the performance (detection and recognition accuracy, robustness of de-
tection) and offer the hope of transferring learned knowledge from one
domain to another.
• Improved speed and accuracy of ISR and C2 decision-making: The auto-
mation of collection, networked dissemination will enable automated
translation of data to a more comprehensive assessment of the naval
situation for commanders in the C2 process. Emphasizing naval tactics,
Hughes and Girrier noted, “AI, as a man-machine enhancing system,
will help commands and staffs keep up with the revolution in data gath-
ering and defend themselves against the growing variety of ever-hotter
new missiles” [5].
 The Role of AI, Automation, and Autonomy 157

The effect of potential AI applications has even brought the introduction

of the term algorithmic warfare to emphasize the significance of the competi-
tion for effective AI algorithms and the impact they will have in future combat
[6]. The competition also includes the development of adversary AI techniques
to deceive or disrupt AI by exploiting the vulnerabilities of their training limi-
tations, imperfect models, and decision processes that cannot fully consider
context.

8.1 Automating Naval Systems

Automation and autonomy are accomplished by the implementing functions
of AI technologies. Here we specifically describe the elements of AI capabilities
that are integrated into ISR sensor platforms (unmanned vessels and vehicles),
sensing systems, networks, and ISR fusion systems.
The elements of AI applied to automation in most naval systems discussed
in this book are summarized in Table 8.1, which is organized by the forward
path of well-known OODA loop elements used to model ISR and C2 func-
tions [7]. The table distinguishes the functions from sensing and representing
knowledge through learning and reasoning, to planning to conduct sensing and
information collection actions. Each column identifies the functional elements
(or software technologies) that are included in the AI umbrella [8].
Sensing. Physical sensing provides quantitative measurements of physical
phenomena (visible or thermal emissions, radar, signal intercepts, biometric
measurements, etc.) to provide computer vision, signature detection (physical
objects in GEOINT and signal objects in SIGINT) and classification. Cyber
sensors monitor network traffic, computer activity, and human-created content
(e.g., public media, social media posts, and other PAI as well as intercepted pri-
vate email traffic). The detected and classified objects produced by sensors are
reported and stored as objects in time, or in maritime parlance, POSITs with
attributes (physical object type, home port, duration at sea, etc.).
Knowledge representation. The knowledge in a system must be represented
in an explicit form, so it can be organized, indexed, and retrieved for use. In
the earliest forms of AI (expert systems) knowledge was stored symbolically;
subject matter experts translated their knowledge into logical rules and data
sets that could detect conditions, execute rules, request more data, conduct
tests, and make recommendations and decisions. Though relatively narrow in
their performance and brittle (unable to adapt) these representations enabled
the first wave of AI (approximately 1980–2000) and introduced the need to
elicit and capture human expertise in rules and computational models. The
next wave brought semantic formalisms to represent static knowledge by ex-
plicit structures (graphs and semantic languages such as the Web Ontology
158 Naval ISR Fusion Principles, Operations, and Technologies

Table 8.1
Categories of Major AI Functions Related to the ISR OODA Loop
Sensing and
Knowledge Planning,
Representation Machine Learning Machine Reasoning Scheduling, Action

·· Sensing (physical, ·· Learning and model ·· Machine reasoning ·· Autonomous systems

cyber, social) and creation: human- (induction, deduction, and agent goal-
human-sourced based learning, abduction); directed control;
information; machine-based ·· Symbolic reasoning; ·· Multiagent systems;
·· Sensor signal learning; ·· First and higher-order ·· Expert system,
detection and ·· Unsupervised logics; production rules;
processing; clustering; ·· Probabilistic ·· Problem solving;
·· Human learned and ·· Supervised ML and reasoning; ·· Search: dynamic
machine-learned unsupervised ML and ·· Bayesian causal programming,
models; clustering; reasoning; combinatorial
·· Ontologies and ·· Regression and ·· Spatial reasoning; optimization, genetic
semantically estimation (linear, ·· Natural language programming;
structured logistic regression); processing and ·· Optimization.
knowledge; ·· Classification and reasoning;
·· Model representation classifiers (simple ·· Hybrid reasoning;
categories: classifiers, decision ·· Explainable
·· Semantic; trees, Bayesian, reasoning;
·· Temporal; classification by ·· Reasoning to
·· Geospatial; regression); decision analysis and
·· Relational graph; ·· Neural network decision-making.
·· Causal graph classification,
(Bayes); deep convolutional
·· System dynamics; networks;
·· Discrete event; ·· Reinforcement
·· Markov model. learning support-
vector machines;
·· Principal component
analysis;
·· Training, testing, and
cross-validation of
ML models.

Language OWL) and ontologies that represented concepts and relationships.

The semantic representation allowed machines to reason (compute inferences,
compare incoming data to stored knowledge) and make decisions based on
knowledge in basic ontologies (fundamental knowledge), domain ontologies
(accumulated knowledge on a specific domain, such as maritime operations),
and current situations that are also represented semantically.
Machine learning. The next wave of AI (beginning in 2000) introduced
the methods of ML that induce computational models from many examples
(empirical data). For example, presented with many examples of overhead ship
 The Role of AI, Automation, and Autonomy 159

images, maritime shipping routes, loitering patterns, and more, the ML algo-
rithm can identify and classify a specific instance based on the learned char-
acteristics of the many examples. This was enabled by significantly improved
computation that enable neural networks to be massively expanded in nodes
and connections to represent classes or patterns in data.
Machine reasoning. Using sensed information about a current situation,
and a priori knowledge about the world (maritime routes, ports, adversary and
commercial ships, normal situations, threats, etc.) a machine can reason about
(or infer) the implication of the situation. The reasoning may be: (1) deduc-
tive—inferring a specific case (i.e., ships and aircraft converging on a target)
from a set of certain conditions (ship and aircraft tracks), or (2) abductive (rea-
soning to the best explanation of uncertain data) using probabilistic or other
uncertain representations.
Planning, scheduling, and action. Planning involves the representation of
environments, goals, constraints, objective (or utility) functions, and actions.
The planner computes the space of alternative actions and the relative utility
of each, then reasons about the effects and utility of those actions to choose
the optimal plan (or set of plans, since several could be of similar utility) and
schedule or sequence of actions that achieves the highest utility value. Planning
can be represented as a graph search problem where the graph nodes are system
states (current state, desired end state, and intermediate states) and the arcs are
actions. Each action has a cost and each path from desired to end state has a
plan cost and utility (for example, the time to archive the goal). The plan solu-
tion is the best path from the initial state to the end state that satisfies the goal
under constraints. The machine planner applies techniques to efficiently search
the space of a possible plan’s schedules. Even a simple plan problem, such as
choosing among seabed-to-space sensors (or sets of sensors) to select a best sens-
ing sequence to track a maritime target, creates a vast search space. Representa-
tion and search methods seek to apply compact representations and efficient
search methods. The ability to integrate these capabilities provide automation
for a range of naval applications (Table 8.2) that increase the volume, coverage,
speed, and completeness of ISR reporting.
While many of the mathematical and software functions listed earlier in
Table 8.1 have existed for decades (not under the label AI), it is the integration
of these capabilities into a learning, reasoning, planning, and adaptive system
that earns the status of an AI system. In maritime ISR and C2 systems, it is the
ability to automate the process to sense (observe), perceive and reason to assess
the situation (orient), conceive and weigh alternatives in the context of naval
tactics (decide), and then to execute C2 (act).
An example of such an automated ISR process was illustrated by the U.S.
National Reconnaissance Office (Figure 8.1) where an ISR collection system
(in this case a set of MultiINT satellites, as described in Chapter 9) is given a
160 Naval ISR Fusion Principles, Operations, and Technologies

Table 8.2
Automation Opportunities in Naval ISR and C2
Intelligence Example AI Automation Opportunities
Area Description/Functions for Maritime Fleet
Intelligence ·· Dynamically translate ·· Translate commander’s intent (qualitative)
strategy-to- intelligence strategy to to dynamic values to prioritize ISR
collection priorities based on changing collection.
planning, maritime situations. ·· Translate priorities to quantitative values to
tasking optimize ISR collection alternatives.
ISR Sensor ·· Dynamically coordinate focus ·· Automate the networking of unmanned
Collection of collection attention on high- collection vehicles (space, air, surface, or
priority emergent events and undersea) and seabed sensor networks.
activities. ·· Automatically coordinate sensing across all
·· Adapt to plan and reschedule domains (and with other elements of joint
an entire network of collection all-domain forces).
systems to reconstitute after ·· Increase autonomy in high-risk, high-payoff
network disruption. automated cyber protection and collection.
·· Optimize or conserve resources ·· Autonomous, miniature robotic systems
based on intelligence value. are being applied in maritime surveillance,
·· Autonomous sensors reconnaissance, and comm relay (air
and actuators capable of vehicles, surface vessels, and undersea
evasion, deception, resource vessels) [9]. Autonomously respond to
optimization, and self- threats to collection systems (e.g., UAVs
preservation. or UUVs).
·· Autonomous surface and seabed sensor
networks report on relevant and anomalous
activities [10].
ISR Processing, ·· Screen and classify massive ·· Screen massive maritime data for naval
Analysis, and data for objects, events, and vessel vs merchant ships and enhance
Dissemination activities. automated ship OBP.
·· Correlate spatially, temporally, ·· Cloud-based correlation of maritime object
or causally related data across reports across all INTs.
all INTs. ·· Cue analysts to suspicious objects, events,
·· Anticipate future activities and activities.
based on ML activity models. ·· Screen open-source media and social
·· Automated natural language media to anticipate civil population
generation and report writing. sentiment, social events, and disruptions
·· Dynamically adapt processing [11].
resources based on current
situation.
C2 for Naval ·· Automate elements of scenario ·· Conduct human-machine assessment,
Operations analysis, planning, and machine generates potential adversary
decision-making. maneuvers;
·· Automate continuous real-time ·· Human-machine decision-making support;
contingency assessment. ·· Machine-based counterdeception cueing
to adversary AI, EW, or other deception
moves.

problem (a situation requiring surveillance) that is parsed into a set of targets

that are used by tasking agents to orchestrate the collection by diverse sensors.
The collected MultiINT data are processed to fuse the information to create
automated reports for analysts and decisionmakers.
 The Role of AI, Automation, and Autonomy 161

Figure 8.1 Elements of a closed-loop MultiINT collection capability [12].

AI systems require a specialized system engineering approach to design,

secure, and deploy them for naval applications [13]. This engineering process
must account for conditions unique to AI functions:

• Management of knowledge in machine-learned models: training, evalu-

ation, update training, transfer learning, and model reuse;
• Model monitoring and logging of data trails and decision sequences for
ongoing evaluation;
• Design for explainability of AI models and processes to allow humans to
understand automated nominations for actions and actions taken;
• Design for security of training processes and training data, model arti-
facts, processes, and protection from adversary AI attacks;
• Assure validation in situations with ground truth and in unexpected
situations outside of boundaries of performance (i.e., verify it will not
err when exposed to situations outside of its limits of knowledge).

In the next two sections, we illustrate specific functions in ISR and C2

that benefit from AI automation implementations.

8.2 Sensemaking in a Naval ISR Context

The term sensemaking refers to the process of gathering and organizing data and
interpreting often incomplete or ambiguous data in context to arrive at a shared
understanding of an emerging situation. The academic study of sensemaking
gained widespread notice when Weick introduced the ideas in the context of
162 Naval ISR Fusion Principles, Operations, and Technologies

psychology and management science in the early 1990s [14], and has now ex-
panded to fields as diverse as business, law, medicine, law enforcement, and
of course military and intelligence analysis. Sensemaking goes beyond the de-
velopment of situational awareness and seeks to inform the selection of COAs
which may be taken in response to the developing situation.
In a naval warfare context, sensemaking involves correctly assessing the
situation both afloat and ashore in the region of interest. The data considered
could include object location and identity obtained by two categories of data:

• Hard data provided by quantitative sensor measurements of physical

phenomena (e.g., imaging sensors, radars, and sonar);

• Soft data that is qualitative such as ships rosters, HUMINT reporting,

and OSINT such as crew social media posts or local news reporting.

The timeframe for hard and soft data can be quite different—hard sen-
sors operate in the moment, while soft sensors may accumulate data relevant to
a maritime situation over a course of days or even months. Sensemaking must
consider data that informs the location and disposition of friendly and allied
forces; neutral forces must be considered as well as enemy forces. Not surpris-
ingly, it is the disposition of enemy forces, and the feasibility and benefit of
one’s own various COAs that naval officers state they find most beneficial to
sensemaking [15]. The geographic extent of naval engagements can be huge,
which leads to large volumes of data that feed into the sensemaking process. Au-
tomation will be required to organize and analyze the information in a timely
manner. This is also true for relatively small areas where naval activity levels are
high [16].
A traditional view of the ISR data gathering process begins with an under-
standing of the concepts of sensing and discovery as they relate to sensemaking.
By sensing we mean the act of applying a system to an area of physical space
or cyberspace over a period of time to collect information. Discovery involves
sensing an area where the entities in the area are not known to exist. The sens-
ing in this case can reveal new entities—examples might include a search radar
discovering aircraft entering its search area, or a web-scraping tool turning up
several keyword hits on a heretofore unknown website. Sensing can be applied
in a more targeted sense as well, seeking to add to or clarify the status of an
entity or location already discovered.
Any strategy for gathering data for sensemaking must first reconcile with
whether sensemaking is viewed as a top-down process or a bottom-up process.
Some argue that sensemaking begins with data—data is collected, processed,
 The Role of AI, Automation, and Autonomy 163

analyzed, and a situation is inferred [17]. Others argue that sensemaking is an

inherently top-down process, and that analysts and decision makers first con-
struct a model of potential situations and, from there, sensors are brought to
bear to gather evidence that will either support or refute the hypothesis [17].
A more balanced approach is presented in Klein’s data/frame model [18]. The
data/frame model views sensemaking as an iterative process of gathering data
that is initially sought to inform a set of hypotheses, then an evaluation is made
of what data is needed to add more information to support or refute these hy-
potheses. Each hypothesis (e.g., the vessel is a foreign guided-missile cruiser)
can be used to select the essential elements of information (EEIs) to confirm
the hypothesis (in the destroyer case: radar signatures, shape and size, speed,
etc.). Once the EEIs are identified, collection requests can be made to obtain
the EEIs.
Additional data needed can be identified at multiple steps along the pro-
cess from sensing to sensemaking. Entity resolution may require multiple looks
from a radar or camera or a combination of the two. Situation recognition
may require observing entities over a period of time or querying soft data to
seek context. Differentiating between two hypotheses that appear similar but
have vastly different implications may depend on collecting a key piece of data
at a specific time. In a sensemaking model developed by Pirolli and Card, the
collection tasks can be generated in small inner loops (e.g., foraging for infor-
mation about a tracking radar), or at a high-level outer loop involving weighty
policy decisions [19]. In most cases, however, the amount of additional data
that is sought by sensemaking may exceed the collection capacity of the avail-
able systems.
In Table 8.3, we compare the sensemaking models of Pirolli and Card,
and Klein with other similar models to show the similar hierarchical approach
to moving from data to frames (hypotheses) to decision-making. (This may also
be compared to the cognitive model introduced earlier in Chapter 1.)
When sensor resource capacity cannot meet demand, a system of priori-
tizing sensing requests is needed to maximize the value of information available
for sensemaking. There are two primary considerations involved in this calcu-
lus—first, the value that any given piece of data has in confirming or refuting
the hypotheses in play. This value can be computed quantitatively, for instance,
by representing the hypothesis as a graph with entities, events, and locations as
nodes and relationships and activities as edges and assigning qualitative values
and relational rules to the elements of the graph [20]. An example of a quanti-
tative approach would be a Bayes network—a structured, graphical representa-
tion (a net) of probabilistic relationships between several random variables. The
user can represent the prior probabilities of outcomes, conditional probabilities
between variables, and then compute the posterior probabilities of outcomes.
164 Naval ISR Fusion Principles, Operations, and Technologies

Table 8.3
Comparison of Sensemaking Models
Pirolli & Card
Outer Loop Sensemaking Data Fusion Sensemaking Example Methods
Data (Klein), Data Collection, Observation Radar, imaging,
foraging (Pirolli sensing collection schedule
& Card) optimization
Shoebox of Entity resolution Objects resolved Correlation and
potentially where objects association, OBP
relevant data are distinguished
Evidence Track association, Object relations, Kalman filter,
grouping relational graphs multihypothesis
tracking
Frame (Klein), Schema Spatial grid, Situation Concept diagrams,
sensemaking the maritime recognition geospatial, graphs
(Pirolli & Card) situation (relational, entity-
assessment relationship, social
networks), temporal
event sequence,
causal graphs
Hypothesis Alternative Situation projection Storytelling
situations
Reality/policy Search for Collection tasking Multihypothesis Evidence-based
loop (Pirolli & support reasoning reasoning
Card) Revaluate Ideation: The Hypothesis generation
process of idea
creation

Such nets can directly represent the structure of input evidence variables, rela-
tions between evidence and hypotheses (inferences represented as conditional
probabilities), and the output hypotheses [21]. As evidence nodes in a Bayes net
are adjusted, the change in probability of belief for each hypothesis is adjusted.
Similarly, a Dempster-Schaeffer representation of the hypothesis could explic-
itly account for the uncertainty in the situation, which each potential sensing
action could reduce accordingly [22]. The second consideration is the value
that resolving a given situation has in the eyes of the decision makers. This can
often be somewhat subjective, but once high-level objectives are assigned rela-
tive value, it is possible to flow down values to the information related to the
branches of each situation. One formulation of these is the goal lattice approach
proposed by Hintz [22].
In extreme cases, decision makers may find themselves facing a situation
where no existing hypothesis fits the situation that is unfolding. In such cases,
a new hypothesis must be generated. Analysts must often then attempt to orga-
nize data according to different clustering strategies and attempt to discover and
 The Role of AI, Automation, and Autonomy 165

define relationships and causality by using tools such as mind-maps, concept

diagrams, and propositional diagrams [23].

Automating Sensemaking for Naval ISR

As mentioned previously, the amount of data that must be organized for sen-
semaking to represent anything beyond the most trivial naval engagement can
be enormous. For this reason, automation and AI must be employed to achieve
decision superiority. We draw a distinction between automation which simply
mechanizes tasks and has a straightforward process, and AI, which employs rea-
soning heuristics and mathematical models to either support human decision
makers or to make decisions in their stead. As we climb the ladder of sensemak-
ing tasks from sensing and discovery to hypothesis generation, the machine
support needed shifts from predominantly automation to predominantly auto-
mated functions.
In the realm of discovery and sensing, many of the common problems
have been solved. Optimal search strategies were developed during WW2 [24]
and are still valid. Sensor platforms and sensor systems can be automated to
perform optimal searches and, in many cases, can refine and optimize track-
ing and entity resolution. A multifunction radar system may have enough on-
board intelligence to optimally balance its energy resources between search and
track functions, even when multiple entities are held in track. An electro-optical
imaging system with coregistered cameras can superimpose thermal or mul-
tispectral imagery over a high-resolution image and process the fused image
information.
Fusing the information from multiple sensors that are addressing the
same target, as mentioned earlier, aids in entity resolution. Fusing information
about different entities and activities is the domain of situation recognition
and can be performed from the geocentric perspective of ABI [25] or object-
based intelligence. In the latter, an object can have a somewhat abstract mean-
ing, and include organizations such as terrorist groups, or shipping lines. The
principles of data fusion relative to these approaches are well understood [26],
and the application of ML is showing progress towards providing an AI solu-
tion to recognizing patterns of activity associated with individual tracks [27,
28]. Only recently has research on recognizing emerging situations involving
multiple entities begun to occur in a naval context, though a framework has
been proposed and demonstrated [29]. Projecting the situation into the future
is more problematic. One of the domains where research in situation projection
is very active is in developing solutions for self-driving cars [30]. Similar safety
related concerns have led to work on immediate safety of navigation issues in
the maritime domain [31]. Multihypothesis tracking as a means of tracking
single entities is a relatively mature capability, but projecting the development
166 Naval ISR Fusion Principles, Operations, and Technologies

of complex situations, such as an overall naval engagement, is complex and

requires much higher-order simulations (than a kinematic model in a Kalman
filter, for example) to examine alternative futures for comparison with incom-
ing observations.
When decision makers find that the evidence they have is not consistent
with any current hypothesis, then it can literally send analysts back to the draw-
ing board. Drawing diagrams, posting notes, and 3 × 5 cards are often the tools
employed in brainstorming sessions to form hypotheses [32]. The process usu-
ally involves grouping related elements of information into clusters. Of course,
clustering is a very mature aspect of ML, as is link analysis, so there is poten-
tial to leverage AI in the hypothesis generation phase of sensemaking [33, 34].
While AI is currently unable to provide the creative spark or judgment that a
human analyst brings to hypothesizing, machines can certainly aid in organiz-
ing vast amounts of data and presenting it in a manner that supports hypotheses
generation.
The final step in our sensemaking cycle is to come full circle and allow
the hypotheses to inform the additional data needed to continue to refine the
shared understanding of the situation. If valuation is tied to the reduction in
hypotheses uncertainty and to the overall mission goals, then a quantitative
value, or at least a relative ranking can be applied to each sensor request. For
modest numbers of sensors and opportunities, optimization algorithms such
as mixed-integer linear programming (MILP) can be applied [35]. However,
the difficulty of the sensor assignment problem increases exponentially with
the number of sensors and targets. At larger scales, organizational as well as
computational issues arise. Multiple stakeholders compete for the same sensors
and multiple organizations control sensors that could be brought to bear. Some
advocate for centralizing sensor resources under a unified controlling agency,
but no way forward has been proposed [36].
The Australian Defence Science and Technology Group (DST-G) has
demonstrated a maritime human-machine fusion system prototype that pro-
vides a virtual battlespace that presents the results of sensemaking and a vir-
tual adviser avatar that interacts with human analysts that query the avatar for
information (about vessels, their track history, relation to other vessels, etc.).
The system, called consensus, integrates many of the technologies described
earlier—high-level fusion of hard and soft data, semantic reasoning about a
maritime situation, and natural language interaction with human users. For
human-machine interaction (maritime analyst-to-consensus), a controlled nat-
ural language called Lexpresso (limited and formal) is bridged to uncontrolled
natural language to provide an intuitive and natural interface to the machine
through speech or text. The virtual adviser reports on maritime events and situ-
ations and coordinates speech explanations with 3-D visual aids [37].
 The Role of AI, Automation, and Autonomy 167

8.3 C2 Automation
The previous section considered AI applications in ISR systems where collec-
tion efficiency and effectiveness are paramount. We now briefly consider the
application of these methods to command where the commander issues orders
for the purpose of bringing about a particular action and control to organize
and employ forces to achieve the effects of command decisions. Specifically, we
address the automation in C2 systems—the facilities, equipment, communica-
tions, procedures, and personnel essential for a commander to plan, direct, and
control operations of forces pursuant to the missions assigned [38].
NATO explored C2 systems anticipating network centric distribution of
forces and developed a basic reference model of the elements of C2 (Figure 8.2,
adapted for this chapter) [39]. Figure 8.2 highlights the areas that will benefit
from A technologies that may include automated functions and human-ma-
chine automation. The figure distinguishes the fundamental elements in both
ISR sensemaking and C2 decision-making with the areas of automation labeled.
The emphasis in the NATO model was on distinguishing an individual unit’s
awareness and shared awareness, and the value chain measured by the quality of
information provided by ISR (e.g., detection, identification, and tracking) and
the quality of decisions that result from informed decision-making (measured
by effect on naval outcomes).
The model recognized the ISR-C2 control loop could be implemented
on a network, and that introduces three fundamental dimensions that must be
considered:

• Allocation of decision rights (from single centralized decision-making

to fully peer-to-peer);
• Patterns of interaction among the actors (from fully hierarchical to fully
distributed);
• Distribution of information (from tight control to broadly disseminat-
ed).

The implications of these options are significant; traditional C2 is cen-

tralized command, fully hierarchical, and tightly controlled—a small single
design point in a broad option space [40]. Distributed operations may retain
a centralized command but seek distributed control mechanisms enabled by
distributed networked information (a shared awareness) and distributed inter-
actions among subordinate maritime units to coordinate collection (ISR) and
fires (C2). The move away from central command to fully distributed com-
mand at the edge was introduced in five NATO maturity levels of C2 (Table
8.4) [41]. The objective of moving toward distributed operations is to achieve
 168
Naval ISR Fusion Principles, Operations, and Technologies

Figure 8.2 NATO C2 reference model with opportunities for automation. (Figure adapted from [39].)
 The Role of AI, Automation, and Autonomy 169

Table 8.4
Five Conceptual Levels of C2 Distinguished by Degrees of Distribution
Level of Allocation of Patterns of Interaction Distribution of
Maturity Command Among Participants Information
5. Edge C2, full Not explicit, self-Unlimited interaction All receive shared
DMO allocated (emergent,
as required to achieve awareness information,
tailored, dynamic) objective shared effects information
4. Collaborative Collaborative process
Significant Info distributed on
C2, collab DMO and shared plan collaboration areas-
functions
3. Coordinated C2, Coordination process Limited and focused Info distributed on
DMO and linked plans interaction as required coordination areas-
functions
2. Deconflicted C2 Establish Very limited and focused Information distributed to
deconfliction to achieve command constrain and close gaps
constraints
1. Conflicted C2 Centralized, no None, follow command Only organic information
allocation
From: [41].

C2 agility—the ability of a force or collection of entities that are participating

in a complex endeavor and in the context of the capability of the entity or col-
lective to focus and converge. It is the synergistic combination of robustness,
resilience, responsiveness, flexibility, innovation, and adaptation [42]. DMO
implementation requires a C2 design that applies these principles appropriately
to allocate functions that are fully distributed and often autonomous [43].
Edge C2, the most demanding level of maturity in DMO still requires a
unifying commander’s intent—an explicit description of the desired endstate,
constraints (time, resources, etc.), and key tasks. All nodes (e.g., surface vessels,
submarines, unmanned vessels, and vehicles) operate in harmony to dynamical-
ly synchronize (plan, schedule, allocation actions, measure effects, and adapt) to
achieve the intent. The intent must be codified such that it is both human and
machine understandable to enable consensus to be reached in human-machine
teaming, where AI processes augment human analysts, operators, staffs, and
commanders [44]. These human-machine teams will allow decision support
processes to create complex decision alternatives, and simulate the effects of po-
tential actions, providing the commander with insight into the relative benefits/
risks of alternative actions in complex maritime operations [45].
A critical consideration in implementing AI capabilities in ISR (collection
management and sensemaking) and in C2 decision-making must be trustwor-
thy—a psychological state in the mind of the user that accepts vulnerability
to outcomes based on positive expectations of the AI performance. This is ob-
tained by the user from statistical (empirical) and experiential evidence. Com-
manders and their staffs must understand the basis for the AI system reasoning,
170 Naval ISR Fusion Principles, Operations, and Technologies

for example, it must be explainable—able to describe explicitly the rationale

for any ISR sensemaking decision, or any C2 recommended plan, schedule, or
action and the expected outcomes.
Trust and ethical considerations for applying AI to intelligence applica-
tions has been explicitly defined by the U.S. Intelligence Community by six
principles. Applications must: (1) respect the law and act with integrity, (2)
be transparent and accountable, (3) be objective and mitigate bias, (4) temper
technological guidance with human judgment, (5) be secure and resilient, and
(6) be informed by science and technology [46]. Key practices recommended
by the National Security Commission on AIs (NSCAI) to implement these
capabilities are summarized in Table 8.5; the table illustrates specific practices
applied to the maritime ISR and C2 applications [47]. The criteria certainly
support the need to establish trust by the commander and users, but these are
applicable to all automated systems.
Of course, C2 decision support must be trusted by commanders as they
become decision recommendations systems that provide supporting decision
alternatives and simulated outcomes that lead to lethal action. Similarly, au-
tonomous and semiautonomous weapons, controlled by C2 decisions, must
be trusted by those who release them. The U.S. DoD, for example, direct that
“autonomous and semi-autonomous weapon systems shall be designed to allow

Table 8.5
Practices for Representing ISR and C2 Objectives and Tradeoffs
Key Practices Naval Op Application Example
1. Consider and ·· For automated sensemaking, specify and document all operating
document value thresholds that have implications for performance (e.g., precision and
considerations in recall).
AI systems and ·· Specify and document all decision tradeoffs and methods of managing
components. uncertainty (measuring, codifying, and combining uncertainty).
·· Specify and define rationale for all acceptable thresholds for false
positives and false negatives (for detection, identification, and
classification).
2. Consider and ·· For ISR collection optimization processes, particularly specify goals,
document value objective functions, weighting criteria for each, and trades.
considerations in AI ·· For C2 decision-making support, specify and document logical limits
systems that rely on based on disallowed outcomes, and constraints on allowed performance.
representations of Also define and document out-of-bound areas where the automated
objective functions. system cannot perform properly (disallowed operating regimes).
3. Conduct ·· For C2 planning systems, identify context-specific changes in goals or
documentation, objectives that would require a revision of parameters on settings or
reviews, and set limits weightings on factors.
based on disallowed ·· Test sensemaking and C2 systems on specific use cases over a wide
outcomes. range of conditions, including out-of-bound conditions, and adversarial AI
deception conditions.
From: [47].
 The Role of AI, Automation, and Autonomy 171

Table 8.6
Candidates for Automated Processes in the Kandigan Straits Scenario
Day Activity Candidates for Automated Processes
Day 1. Exploit space EO imagery ·· Satellite onboard screening of ship objects and cross-
1 and SAR to locate the Red SAG satellite cueing to enable reacquisition and high revisit
force; track the SAG. key targets.
·· Red ship classification by EO-SAR fusion.
2. Estimate likely target island(s) ·· Using previously machine-learned patterns of behavior of
and landing sites. Direct the Red fleet, staff-machine teams perform C2 planning;
available ISR to identify any Red run exploratory simulations of alternative Red scenarios
mining operations. and alternative Blue COAs.
·· Initiate ISR task planning simulations to prioritize space
and coordinated unmanned autonomous collection
missions.
3. Deploy USVs, UUVs, and ·· Deploy unmanned reconnaissance air vehicles, surface
persistent UAVs and task vessels on autonomous reconnaissance, and relay
submarines to predeployed missions.
seabed sonar sensors. Track ·· Autoallocate seabed-to-space space sensing to track high-
Red SAG, submarines, and air priority vessels, monitor air activity, and missile sites.
support.
4. Monitor civilian channels ·· Using machine-learned RF traffic patterns, autoadapt
in social to estimate popular SIGINT and cyber monitors to exploit reporting and
support/opposition. estimate civilian sentient, and focus SIGINT on Red C2
actions.
5. Receive strategic guidance ·· Staff-machine teams update C2 and ISR planning based on
updates; position for new guidance; run exploratory simulations of alternative
surveillance. Red attack scenarios and effects of alternative Blue COAs.
Day 1. Receive strategic guidance ·· Update persistent space surveillance and air
2 and prepare C2 plans for reconnaissance tasking to assess readiness of shore-
engagement; focus ISR planning based missile sites.
to support cooperative long- ·· Update C2 planning for long-range engagement of Red
range missile engagement SAG.
supported by USVs and UAVs in ·· Reposition UAVs and USVs to preform relay for
terminal phase. cooperative engagement of Red SAGs.
2. Locate Red subsurface ·· Autoplan locations of UUVs to relay acoustic reporting to
threats; employ UUVs and Blue SAG; relay commands to task acoustic sensors to
seabed sensors to locate the autonomously report high-priority subsurface targets.
Red submarine threat. ·· Staff-machine teams simulate Red submarine maneuvers
and reallocate all subsurface sensing.
3. Track Red fleet; task satellite; ·· Staff-machine teams run battle management simulations
manage organic UAVs tracking to select COAs that support special forces operations.
and relay for cooperative ·· Autonomous UAVs and USVs position for strike support
surface-surface missile and endearment relay.
engagements.
4. Monitor Red C2 channels to ·· Plan and initiate coordinated and distributed counter C2
support jamming operations; disruption and deception; monitor effects and continuously
analyze the traffic patterns. Map adapt.
cyber network data to engage
Red cyber networks.
5. Monitor civilian channels; ·· Autoextract, filter major opposition sources and channels;
conduct counterinformation simulate effects of alternative counterinfo campaigns
campaign. and provide to IP Staff to counter misinformation and
opposition sources.
172 Naval ISR Fusion Principles, Operations, and Technologies

commanders and operators to exercise appropriate levels of human judgment

over the use of force” [48].

8.4 Automation in the All-Domain ISR Fusion Scenario

We now return to the Kandigan Straits scenario in Chapters 2 and 3 to illus-
trate how AI, autonomy, and human-machine teaming capabilities in the Blue
fleet might operate in the unfolding crisis. This scenario serves to illustrate the
required functions to achieve the essential efficiency of information collection,
networking to proper nodes, sensemaking, and decision-making to optimize
the ISR and C2 processes. In Table 8.6 we identify the areas in the scenario
where AI and automated functions can play a role in enhancing Blue’s ISR and
C2 agility to respond to the threat across all domains.
The 2-day scenario required a rapid distributed response to an emerging
SAG threatening islands in the straits. Table 8.6 illustrates the role of adaptive
planning by human-machine teams using simulation to explore alternatives,
automated ISR collection (space, air, surface, to seabed) that adjusts to the fast-
paced scenario, use of prior machine-learned knowledge about the red fleet, and
machine reasoning to rapidly screen ISR data and maintain tracks of adversary
forces and their projected actions. Of course, all of this relies on the resilient
and robust network that exchanges data between sensors, platforms, relays, data
fusion nodes, and combatant aircraft and vessels.

Endnotes
[1] For a wider consideration of naval applications, see Heller, C. H. (2019) “The Future
Navy—Near-Term Applications of Artificial Intelligence,” Naval War College Review,
Vol. 72, No. 4, Article 7, 2019, https://digital-commons.usnwc.edu/nwc-review/vol72/
iss4/7.
[2] Terms adapted from Shattuck, L. G., “Transitioning to Autonomy: A Human Systems
Integration Perspective,” Presentation at Transitioning to Autonomy: Changes in the Role
of Humans in Air Transportation, March 11, 2015, p. 5, human-factors.arc.nasa.gov/
workshop/autonomy/download/presentations/Shaddock%20.pdf.
[3] We adopt the very basic Merriam-Webster definition, then provide our own description
of human behaviors. The English Oxford Living Dictionary defines these behaviors as vi-
sual perception, speech recognition, decision-making, and translation between languages.
Different definitions of AI distinguish between whether a capability just thinks rationally,
thinks like a human, or externally appears to act like a human (or is indistinguishable
from a human).
[4] “Summer Study on Autonomy,” Defense Science Board, Office of the Under Secretary of
Defense for Acquisition, Technology and Logistics, Washington, D.C., June 2016.
 The Role of AI, Automation, and Autonomy 173

[5] Hughes, W., and R. Girrier, Fleet Tactics and Naval Operations, Annapolis, MD: Naval
Institute Press, 2018, p. 191.
[6] “Establishment of an Algorithmic Warfare Cross-Functional Team (Project Maven)” Dep-
uty Secretary of Defense, DoD Memorandum, April 26, 2017.
[7] The OODA loop was developed by Col. John Warden and is a high-level model of the
military ISR and C2 loop that considers the human decision-making role and its depen-
dence on observation and orientation (orienting observations in a perceptual framework
for decision-making). The OODA model applies to ISR which provides the observe func-
tion and the C2 process (orientation to the mission, decision analysis, and decision). The
act process includes military actions, as well ISR collection tasking and sensor control. See
Boyd, J. R., “The Essence of Winning and Losing,” unpublished briefing, January 1996.
Also see Hammond, G. T., The Mind of War: John Boyd and American Security, Washing-
ton D.C.: Smithsonian Institute Press, 2001; and Coram, R., Boyd: The Fighter Pilot Who
Changed the Art of War, Boston, MA: Little, Brown, 2002.
[8] There exist many taxonomies of AI technology elements. For a comprehensive comput-
ing taxonomy see the ACM Computing Classification System (https://dl.acm.org/ccs).
See the categories and subcategories of AI topics in the ACM taxonomy at http://aima.
cs.berkeley.edu/topics.html.
[9] Department of the Navy, Unmanned Campaign Framework, March 16, 2021.
[10] Heidelmann, J., M. Stovanovic, and M. Zorz, “Underwater Sensor Networks: Applications,
Advances and Challenges,” Phil. Trans. Royal Society A (2012) 370, pp. 158–175, https://
royalsocietypublishing.org/doi/pdf/10.1098/rsta.2011.0214.
[11] Sathappan, M., et al., “EMBERS at 4 Years: Experiences Operating an Open Source
Indicators Forecasting System,” March 31, 2016, arXiv:1604.00033v1 [cs.CY].
[12] Figure adopted from the NRO Director’s comments at http://geointv.com/archive/geoint-
2013-keynote-betty-j-sapp/, “Future Ground Experiment Called Sentient” section at
12:36 in the video segment. The Figure is adopted from the graphic that Director Sapp is
describing in the transcript. April 2014.
[13] Bosch, J., H. H. Olsson, and I. Cmcovik, “Engineering AI Systems: A Research Agenda,”
June 3, 2020, https://arxiv.org/pdf/2001.07522.pdf.
[14] Weick, K. E., Sensemaking in Organizations, Vol. 3., Sage, 1995.
[15] Waldenström, C., “What is Difficult in Naval Sensemaking,” in Proceedings of the 13th
International Command and Control Research and Technology Symposium, 2008.
[16] Lim, N., and C. De Xian, “Maritime Sense-Making and The Role of Big Data Analytics
for Enhancing Maritime Security,” 2020.
[17] Moon, B. M., and R. R. Hoffman. “How Might ‘Transformational’ Technologies and
Concepts Be Barriers to Sensemaking in Intelligence Analysis,” in Presentation at the
Seventh International Conference on Naturalistic Decision Making, 2005.
[18] Klein, G., J. K. Phillips, E. L. Rall, and D. A. Peluso, “A Data-Frame Theory of
Sensemaking,” in Expertise Out of Context: Proceedings of the Sixth International Conference
on Naturalistic Decision Making, New York: Lawrence Erlbaum Associates, Vol. 113, 2007.
174 Naval ISR Fusion Principles, Operations, and Technologies

[19] Pirolli, P., and S. K. Card, “The Sensemaking Process and Leverage Points for Analyst
Technology,” paper presented at the International Conference on Intelligence Analysis,
McLean, VA, May 2005.
[20] Tecuci, G., D. Marcu, M. Boicu, and L. Kaiser, “Instructing a Cognitive Agent to Perform
Sensemaking in Intelligence, Surveillance and Reconnaissance,” in Proceedings of the Eight
Annual Conference on Advances in Cognitive Systems, 2020, pp. 10–12.
[21] The Bayes Network is a fundamental mechanism for probabilistic inference, is a mature
mathematical method, and has been employed since the earliest days of expert AI
system. For classic introduction, see J. Pearl, Probabilistic Reasoning in Intelligent Systems,
Burlington, MA: Morgan Kaufmann, 1998.
[22] Hintz, K. J., Sensor Management in ISR, Norwood, MA: Artech House, 2020.
[23] Hoffman, Robert R., Tom Eskridge, Simon Henderson, Jonathan Jenkins, and Brian
Moon. “Propositional Diagrams for Intelligence Sensemaking: Examples and Case
Studies,” American Intelligence Journal, Vol. 32, No. 1, 2015, pp. 122–135.
[24] Koopman, Bernard O, “A Theoretical Basis for Method of Search and Screening,”
Columbia University, New York, 1946.
[25] Biltgen, P., and S. Ryan, Activity-Based Intelligence: Principles and Applications, Norwood,
MA: Artech House, 2016.
[26] Antony, R. T., Data Fusion Support to Activity-Based Intelligence, Norwood, MA: Artech
House, 2015.
[27] Pitsikalis, M., and A. Artikis, “Composite Maritime Event Recognition,” in Guide to
Maritime Informatics, Springer, Cham, 2021, pp. 233–260.
[28] Gingrass, C., D. I. Singham, and M. P. Atkinson. “Shape Analysis of Flight Trajectories
Using Neural Networks,” Journal of Aerospace Information Systems, Vol. 18, No. 11, 2021,
pp. 762–773.
[29] Vouros, G. A., et al. “Increasing Maritime Situation Awareness Via Trajectory Detection,
Enrichment and Recognition of Events,” in International Symposium on Web and Wireless
Geographical Information Systems, Springer, Cham, 2018, pp. 130–140.
[30] Casas, S., C. Gulino, R. Liao, and R. Urtasun. “SpAGNN: Spatially-Aware Graph Neural
Networks for Relational Behavior Forecasting from Sensor Data,” in IEEE International
Conference on Robotics and Automation (ICRA), 2020, pp. 9491–9497.
[31] Marié, V., I. Bechar, and F. Bouchara, “Real-time maritime situation awareness based on
deep learning with dynamic anchors,” in 15th IEEE International Conference on Advanced
Video and Signal Based Surveillance (AVSS), 2018, pp. 1–6.
[32] Hoffman, R. R., T. Eskridge, S. Henderson, J. Jenkins, and B. Moon. “Propositional
Diagrams for Intelligence Sensemaking: Examples and Case Studies,” American Intelligence
Journal, Vol. 32, No. 1, 2015, pp. 122–135.
[33] Chan, J., S. Dang, and S. P. Dow. “Comparing Different Sensemaking Approaches for
Large-Scale Ideation,” in Proceedings of the 2016 CHI Conference on Human Factors in
Computing Systems, 2016, pp. 2717–2728.
[34] Ancona, D., “Framing and Acting in the Unknown,” S. Snook, N. Nohria, & R. Khurana
(eds), The Handbook for Teaching Leadership, 2012, pp. 3–19.
 The Role of AI, Automation, and Autonomy 175

[35] Richards, J., A. Patel, A. Thorpe, and R. Schlossman, “Autonomous Multi-Platform

Sensor Scheduling for Intelligence Surveillance and Reconnaissance,” No. SAND2019-
11381C, Albuquerque, NM: Sandia National Lab (SNL-NM), 2019.
[36] Katz, B., “The Intelligence Edge: Opportunities and Challenges from Emerging
Technologies for US Intelligence,” 2020.
[37] Lambert, D. A., A. Saulwick, and K. Trentelman, “Consensus: A Comprehensive Solution
to the Grand Challenges of Information Fusion,” Proc. 18th International Conference on
Information Fusion, Washington, D.C., July 6–9, 2015.
[38] Joint Publication (JP 6-0) U.S. Joint Staff.
[39] NATO System Analysis and Studies Panel 050, Exploring New Command and Control
Concepts and Capabilities, AC/323(SAS-050)TP/50, January 2006, http://www.dodccrp.
org/files/SAS-050%20Final%20Report.pdf.
[40] For the classic naval C2 model, see: Lawson, J. S., Jr. “Naval Tactical C3 Architecture,
1985–1995,” Signal 33:10, August 1979, pp. 71–79 and Lawson, J. S., Jr., “Command
and Control as a Process,” IEEE Control Systems Magazine, Vol. 16, March 1981, p. 5.
For newer views, see Alberts, D. S., and R. E. Hayes, “Understanding Command and
Control,” Command and Control Research Program CCRP, 2006 and Pigeau, R., and
Carol McCann, “Reconceptualizing Command and Control,” Canadian Military Journal,
Spring 2002, pp. 53–63.
[41] Alberts, D. S., R. K. Huber, and J. Moffat, “NATO NEC C2 Maturity Model,” Command
and Control Research Program (CCRP), February 2010.
[42] Ibid, p. 38
[43] For a Navy study focused on DMO, see Corbett, L, et al., “Command and Control for
Distributed Lethality,” Naval Postgraduate School, June 2017. For an Air Force view of
centralized command, distributed control, and decentralized execution, see S. Mulgund,
“Evolving the Command and Control of Airpower,” U.S. Air Force: Air University,
April 21, 2021, https://www.airuniversity.af.edu/Wild-Blue-Yonder/Article-Display/
Article/2575321/evolving-the-command-and-control-of-airpower/.
[44] “Human-AI Teaming: State of the Art and Research Needs,” Washington, D.C.: The
National Academies Press, National Academies of Sciences, Engineering, and Medicine
2021, https://doi.org/10.17226/26355.
[45] Waltz, E., “Situation Analysis and Collaborative Planning for Complex Operations,” in
Proceedings of the 13th International Command and Control Research Symposium, Office
of Assistant Secretary of Defense, Bellevue, WA, 2008. See also Waltz, E., “Modeling
the Dynamics of Counterinsurgency (COIN),” in Analytical Tools for Irregular Warfare,
Proceedings of NATO System Analysis and Studies Panel Specialists Meeting (SAS-071/RSM-
003), Ottobrunn, Germany: NATO Research and Technology Organization, 2009.
[46] See “Principles of Artificial Intelligence Ethics for the Intelligence Community,” Office
of the Director of National Intelligence, https://www.dni.gov/index.php/features/2763-
principles-of-artificialintelligence-ethics-for-the-intelligence-community. See also
“AI Principles: Recommendations on the Ethical Use of Artificial Intelligence by the
Department of Defense: Supporting Document,” DoD Defense Innovation Board,
October 31, 2019, https://media.defense.gov/2019/Oct/31/2002204459/-1/-1/0/DIB_
AI_PRINCIPLES_SUPPORTING_DOCUMENT.PDF. See also the Department of
176 Naval ISR Fusion Principles, Operations, and Technologies

Defense Law of War Manual, U.S. Department of Defense (Dec. 2016), https://dod.
defense.gov/Portals/1/Documents/pubs/DoD%20Law%20of%20War%20Manual%20
-%20June%202015%20Updated%20Dec%202016.pdf?ver=2016-12-13-172036-190.
[47] “Key Considerations for the Responsible Development and Fielding of Artificial
Intelligence,” National Security Commission on Artificial Intelligence (NSCAI), April 26,
2021.
[48] Department of Defense Directive 3000.09 on Autonomy in Weapons Systems, U.S.
Department of Defense November 21, 2012, p. 2, https://www.esd.whs.mil/Portals/54/
Documents/DD/issuances/dodd/300009p.pdf.
9
Distributed Space Maritime Surveillance
The importance of space in naval reconnaissance was recognized early in the
twentieth century. Aerial surveillance had proved its utility as early as the Amer-
ican Civil War via the use of manned lighter-than-air balloons and continued
to mature as powered aircraft took on patrol roles during World War I, the
interwar years, and World War II. Indeed, aerial reconnaissance was often the
decisive factor in the carrier battles in the Pacific Theater, the Battle of Mid-
way being a prime example. Although the technology to gain a foothold in
space would not be available until the mid–twentieth century, writers had been
speculating about it in the nineteenth century and fanciful and serious science
fiction addressed the topic throughout the early twentieth century. It is no sur-
prise then that strategic thinkers in the U.S. Navy were poised to act when space
technology came within reach shortly after World War II, forming a Commit-
tee for Evaluating the Feasibility of Space Rocketry on October 9, 1945 [1].
Throughout the late 1940s, the U.S. Navy continued to develop con-
cepts for reconnaissance and communication relay satellites. By 1955 the White
House announced that the United States would launch an Earth-orbiting satel-
lite in the International Geophysical Year 1957–1958. Much to the embarrass-
ment of the United States, the Soviet Union was first to space with Sputnik in
October 1957. The effect was to prompt U.S. politicians to accelerate the U.S.
space programs, both military and civilian [2]. In 1958 the Navy conceived of
using a satellite to “raise the periscope” of a submarine-based ELINT system by
placing it on a satellite. The resulting program, DYNO, was a highly classified
program at the time, and was given cover as an unclassified scientific research
mission known as GRAB (Galactic Radiation and Background). The GRAB/
DYNO satellites launched beginning in June 1960 and were capable of collect-
ing ELINT on Soviet intercontinental ballistic missiles (ICMB) early warning

177
178 Naval ISR Fusion Principles, Operations, and Technologies

radars, making them the world’s first reconnaissance satellites. Although the
program was short-lived (ending in 1962), it was followed by a similar program
(POPPY), and by the mid-1960s satellite reconnaissance had become a perma-
nent strategic priority [3].
The vital need for space-based photographic reconnaissance was made
evident when Gary Powers’ U-2 reconnaissance aircraft was shot down over the
Soviet Union in May 1960. Manned overflights were discontinued, and the
launch of the CORONA satellite system only 3 months later became the only
realistic means of monitoring the USSR’s nuclear arsenal [3]. Even with early
photoreconnaissance technology (40-ft spatial resolution and a requirement to
drop film canisters to a recovery aircraft) CORONA proved how invaluable sat-
ellite reconnaissance could be in surveilling targets deep within denied enemy
territory [4].
By the early 2000s most modern naval forces have become reliant on their
own space assets or leveraged assets for communications, C2, ISR, navigation,
and METOC services. Furthermore, even nations that have invested heavily in
space technology must continually upgrade and replace aging systems, a task
that has traditionally been challenging considering the multi–billion dollar
costs and multiyear design times that have until the last 5-10 years been typi-
cal for satellite systems. In 2005 the U.S. Navy articulated its needs for space
capabilities as shown in Table 9.1. These needs are expected to continue to be
relevant for many decades into the future.

Table 9.1
Navy Space Mission Areas
Capabilities Required for Navy Strike
Space Mission Area Missions
ISR: Responsive persistent ISR (imagery and signals) Target identification, location, and tracking,
of static and moving targets ELINT-enabled target identification
Meteorology and oceanography: Continuous tactical Steaming direction for carrier aircraft
weather prediction, cue appropriate ISR assets launch, maximum steaming speed
depending upon acoustic/thermal modeling of the
littorals, weather
Communications: All information must be moved, Air tasking orders, precision weapon target
often at BLOS ranges, sensor-analysis-decisions- coordinates, intelligence coordination,
shooter-weapon links mission plans
Theater and ballistic missile defense: Defense- Fleet defense rom ballistic and hypersonic
enabled freedom of maneuver missiles
Position, navigation, and timing: Global Positioning Precision target location, GPS-guided
System (GPS) location information, GPS-timing munitions
enables communications coordination
Space control: Ensure access to national space assets Information operations via space links
and provide protection from detection by hostile enabled
assets
From: [5].
 Distributed Space Maritime Surveillance 179

Throughout the Cold War and into the early twenty-first century, the
U.S. Navy led the world in application of undersea and space technology to
surveillance in the maritime domain. However, in the last decade changes in the
economics of Earth observation services have enabled other nations to have ac-
cess to improved maritime surveillance. The current availability of commercial
spacecraft individually and in constellations for all international navies increases
the contribution of space collection and communication to all international na-
vies. This capability is very significant and is a disruptive technology, enabling
new and rapid ISR capabilities and rendering the oceans nearly transparent
under some circumstances. This chapter introduces the disruption that large
sensing and communication constellations enable.

9.1 Smallsat Constellations for Sensing and Communication

Small satellites began as tools for academic experimentation and in lab projects
to teach development and launch of very small satellites [6]. It was possible for
students to design and assemble functional satellites from 3-D printed parts
and commercial, off-the-shelf (COTS) electronics, and leverage ride-share op-
portunities for launch. Standards were quickly adopted with a satellite with a
volume of 10 cm per side cube and 1 kg weight being defined as a one unit, or
1U, vehicle. As 1U, 3U, and 6U designs became common, commercial com-
panies began offering standard busses for satellites with these form factors [7].
At the same time, commercial orbital transportation services to launch these
satellites emerged and enabled the cost-effective launch of large constellations
of smallsat spacecraft [8]. As a result, CubeSats have now become serious tools
for a wide variety of Earth observation missions [9] including maritime domain
awareness [10].
The low cost and short development timeframes of smallsats has been
game changing. The short time from design to deployment (usually 6 to 18
months) means that smallsats are deployed with state-of-the-art sensors and
electronics. The low cost means that very large constellations can be deployed
for the cost of a single legacy satellite. The main advantage of large constella-
tions is that they provide a high revisit rate. Consider the fact that the Planet’s
Wave Flock Constellation advertises the ability to deliver up to 12 daily revisits
for virtually any point on Earth [11]. Compare this to one of the earliest and
the longest continuously running commercial imagery programs using large
satellites, Landsat [12], which only revisited the same spot on Earth once every
16 days [13]. Consider the capability of some of the pioneering smallsat con-
stellations that illustrate the commercial approach to terrestrial missions that
are like those desired to meet the naval ISR missions in Table 9.1.
180 Naval ISR Fusion Principles, Operations, and Technologies

Revisit rate is a critical factor to maintain custody of dynamic naval tar-

gets; increasing collection rate is the increase in the number of vehicles in a
constellation, and smallsat technology has significantly increased the feasibility
and has demonstrated the operation of large constellations that can increase
EO and radar observations rates from hours to tens of minutes. The disrup-
tive implications of large smallsat constellations are significant as they provide
greatly expanded Earth observation (for naval ocean surveillance) and com-
munication network capacity (for naval networking) [14]. The U.S. DoD has
publicly recognized the value of adopting multiple large and diverse commer-
cial constellations to enable increased defense collection [15]. A small sample of
U.S. companies expanding their constellations of commercial remote sensing
offerings are provided in Table 9.2.
We consider the potential of these commercial satellites to support the
naval ISR mission by illustrating three examples of large-scale smallsat constel-
lations for sensing and look at the resources required to provide worldwide daily
coverage in visible imagery, SAR, and ELINT/signal intelligence.

9.1.1 Visible Imagery: Planet Labs Flock

Planet Labs states that the mission of their company is “to image the entire
Earth every day and make global change visible, accessible, and actionable”
[17]. Their approach has been two-fold: (1) launch a constellation of 3U Cube-

Table 9.2
2022 Major U.S. Commercial Earth Observation Constellations (At the Time of This Book)
Constellation Approximate Revisit Rate of Full
Vendor Description Constellation
Planet Dove EO more than 150 Dove 3U CubeSats Up to 12 visits per day (daytime
at medium res, and 21 SkySats at imagery) or once per hour
high-resolution. Dove provides daily
imaging of the Earth
Blacksky EO 24 satellites inclined orbit full Peak of 15 hourly visits per day over
constellation certain locations with 12 satellites
Capella SAR full constellation 36 vehicles Less than 1 hour revisit with full
constellation
HawkEye 360 RF emissions location by a cluster Average global revisit rate will be
of three spacecraft that geolocate less than 1 hour
signals: Plan 7 3-satellite clusters
Maxar Maxar WorldView-1, -2,-3 and GeoEye-1 Average revisit time of <1 day
pan and multispectral imagers, (WorldView-3)
imaging 5-cm class resolution
Legion EO constellation 6 satellite, Up to 15 revisits per day
planned operations begin with first
2 satellites in 2023 [16] 30-cm class
resolution
 Distributed Space Maritime Surveillance 181

Sats called Doves, weighing only 11 lbs, and (2) augment the low-resolution
Dove constellation with a smaller number of high-resolution SkySat smallsats
(purchased from Google, who launched them with the name SkyBox). The
SkySat craft are still small, but much larger than the Dove’s, measuring 80
cm in length and weighing 220 lbs [18]. Most of the mission use cases that
Planet is marketing do not require exquisite spatial resolution. For municipal
planning, energy use prediction, agricultural monitoring, and mapping mining
activity, a resolution of 1m to 3m is sufficient to estimate the size of areas being
monitored.
Daily coverage is probably adequate for frequency of industrial/agricul-
tural updates, but for other use cases, such as naval fleet monitoring and disaster
management where wildfires and floodwaters need to be tracked, daily coverage
is barely enough. Planet seems to have made a compromise here. No doubt, this
is tied to the goal of providing daily change detection. For visible and thermal
imaging applications, changes in time of day can make significant differences
in the image interpretation due to changes in shadow length, contrast, reflec-
tivity, and thermal variations. For this reason, it is desirable to perform change
detection on images taken at the same time of day. Planet accordingly chose a
450-km high sun-synchronous orbit (SSO) at a 97-deg inclination for the set of
spacecrafts known as the Flock 3p constellation (Doves fly in Flocks).
Finally, though most customers would be satisfied with visible-band im-
agery, the agricultural customers need more information. For monitoring crop
health, it is essential to also have the NIR spectrum available, so Planet has se-
lected sensors that enable them to capture the four bands in the visible-to-NIR
spectral range in a single camera with a 90-mm aperture, fixed pointing straight
down (nadir).
In order to meet the daily revisit requirement, the constellation design
called for 150 satellites in SSO, each taking an image once per second. Planet
accomplished this over several launches, the largest launch taking place on Feb-
ruary 14, 2017 when the Flock 3p constellation of 88 Dove 3U CubeSats was
launched on a single vehicle, India’s ISRO PSLV-C37. Planet currently has over
180 satellites in operation and their products generated $113M in revenue in
2020 while they estimate a future market potential of $75B [19].

9.1.2 SAR: Capella

Capella Space Corporation, or Capella, is a satellite imaging company that
provides all-weather imaging capability from a constellation of smallsats us-
ing SAR—an active imaging technique that utilizes radar antennas that emit
pulses of microwave radiation at a pulse repetition frequency of thousands of
pulses per second. The radar receiver measures the time, strength, and phase
characteristics of the pulse echoes and uses this data to generate 2-D images
182 Naval ISR Fusion Principles, Operations, and Technologies

[20]. The X-band radar used by Capella can provide resolution as fine as 35 cm
in high-res spot mode. There are also sliding spotlight (50-cm resolution) and
stripmap (2m resolution) modes [21]. While Capella can reach a good portion
of the Earth’s surface, the 220-lb smallsat design is somewhat limited in power
and can only provide 9 minutes of collection each orbit [21].
Capella’s initial commercial constellation, known as Sequoia, had six sat-
ellites in a 535-km, 45-degree inclination orbit, with a total of 40 satellites
planned. This will allow customers to schedule collections from 1 to 7 days
in advance. Sequoia checks for schedule updates every 15 min but maintains
a strict “no-bump” policy, manning that scheduled tasks will not be cancelled
to accommodate newer requests [22]. The 40 satellites will provide daily revisit
rates in the coverage area. Note that daily coverage is obtained with far fewer
satellites than required for the Planet Labs constellation. There are two factors
contributing to this: (1) a smaller coverage area, extending only to −48.9°S to
+48.9°N, and (2) the ability to point the sensor 45 degrees from nadir [21],
compared to Dove’s 97-degree angle of inclination and fixed nadir sensor.
It is important to notice that the Dove and Capella imaging capabilities
are quite different. The Dove SSO and fixed nadir-pointing sensor is purpose-
fully designed to achieve a change detection capability, comparing identical
day-day images. Dove was never designed for target surveillance of the open
ocean, though its imaging does collect images that provide vessel-detection
quality products. On the other hand, Capella offers taskable SAR collection
which seems of potential utility to maritime surveillance mission. Several fac-
tors that distinguish the utility of SAR versus EO sensors include the fact that
SAR does not require separate illumination and can therefore create images at
night and through weather; EO imagery is daytime only and is occluded by
clouds. We use these commercial imaging satellites for illustrative purposes, and
do not imply that in their current form they are immediately suitable for the
maritime mission.

9.1.3 RF Signals Collection: HawkEye 360

The final commercial constellation we will consider is the HawkEye 360 con-
stellation of RF monitoring satellites. The receivers on the HawkEye satellites
can detect maritime navigation radars, maritime location reporting beacons,
emergency beacons, and VHF and UHF push-to-talk radios [23]. The antici-
pated market is RF intelligence for defense, maritime awareness for commercial
and naval customers, and monitoring and aiding natural disaster recovery op-
erations [24].
In the maritime domain, RF monitoring of navigation radars and AIS
beacons can provide real-time situational awareness and pattern-of-life map-
ping for naval commanders, as well as commercial maritime customers. For
 Distributed Space Maritime Surveillance 183

example, AIS is a required navigation signal on all vessels above a given size.
AIS was originally intended for use from ship-to-shore and ship-to-ship use for
safety of navigation [25]. In this capacity, AIS is limited to a range of about 20
nautical miles. With the capability of monitoring AIS continually from space,
nefarious activities such as illegal fishing, smuggling, and even piracy can be
detected. For example, analytics have been developed to detect illegal fishing
based on the tendency of vessels to turn off their AIS beacons when approach-
ing restricted fishing areas [26, 27]. Detecting these activities is only practical
when space-based observation makes it possible to continually observe patterns
of vessel activity in the open ocean, far beyond the range of shore-based sensors.
The HawkEye constellation will consist of seven clusters of three satellites
per cluster and will provide a revisit time of only 40 minutes virtually anywhere
on the Earth. Three clusters were launched by the end of 2021 [28]. The 15-kg
satellites measure 40 × 27 × 20 cm. By operating in a cluster of three, they can
accurately geolocate signals to within a few hundred meters [29].

9.2 A Networked Constellation for Ocean Surveillance and ISR

The implications of integrating large constellation observation and communi-
cation satellites for maritime surveillance is significant:

• The ability to maintain near-continuous global ocean surface surveil-

lance;
• The ability to communicate satellite observations of ships, ports, and
naval facilities to other satellites, to naval fleets below, and to ground
stations in near real-time;
• The ability for one satellite to dynamically tip other satellites to search
or focus on naval targets detected by the first satellite. Of course, this
presumes the ability of an imaging satellite to point off-nadir to a loca-
tion provided in the tip. Note that this could be enabled by a satellite
relay transport layer that is described in the next paragraphs;
• The ability for surface vessels to intercommunicate over a high-band-
width network (data, C2, voice, video).

Large constellations are a very disruptive technology as these capabilities

are implemented and delivered to navies worldwide. To achieve a useful capa-
bility for naval ISR applications from these commercial observation satellites,
the data must be able to detect and classify ocean surface objects with high
resolution (to classify and identify ships) and a very low latency (on the order of
minutes from collection to delivery to a user or data fusion node). This can be
184 Naval ISR Fusion Principles, Operations, and Technologies

achieved by immediate relay of the acquired data through a transport system to

other satellites or to ground processors.
The DoD is planning such an integrated observation approach using the
U.S. Space Development Agency (SDA) transport layer of spacecraft to relay
data from sensing satellites to ground or to other sensing satellites [30]. The
transport layer is a network of data-relay satellites in LEO (as many as 150 sat-
ellites in orbit) that enable spacecraft to relay data: (1) from spacecraft to relay,
then to ground, and (2) from spacecraft to relay to another spacecraft to tip
areas for further collection.
Figure 9.1 shows the SDA architecture planned by the U.S. and includes
the following components [31, 32]:

• A space transport layer is a global mesh network providing 24/7 data and
communications. The key technology required is an optical intersatel-
lite link (OISL) to provide reconfigurable, multiprotocol, intersatellite
optical communications with the ability to connect heterogeneous con-
stellations that operate on different OISL specifications, operating on
platforms in LEO.
• A tracking layer: Provides tracking, targeting, and advanced warning of
missile threats.
• A custody layer: Provides “all-weather custody of all identified time-crit-
ical targets” [33].
• A deterrence layer: Provides space situational awareness—detecting and
tracking objects in space to help satellites avoid collisions.

Figure 9.1 SDA National Defense Space Architecture (NDSA) [35].

 Distributed Space Maritime Surveillance 185

• A navigation layer: The constellation can also provide alternative posi-

tioning, navigation, and timing (PNT) services in case GPS is blocked
or unavailable.
• A battle management layer: A command, control, and communications
network augmented by AI that provides self-tasking, self-prioritization,
onboard processing, and dissemination.
• A support layer: Ground C2 facilities and user terminals, as well as rapid-
response launch services.

The relay (transport) constellation and commercial sensing spacecraft

(EO, SAR) can be controlled by issuing collection requests on specific areas,
and spacecraft-to-spacecraft cueing requests may be passed through the relay
network. This technology will enable higher revisit rates to adaptively request
collection from all available assets (government and commercial) during periods
of demand when dynamic targets are on the move or conducting activities that
must be monitored [34].

9.3 Contributions to Naval ISR

A common characteristic of any conventional, terrestrial-based surveillance sys-
tem is that the operational range of the platform is limited by fuel and limited
field of view. How limited? Let’s consider the state-of-the-art in long range
persistent platforms—the Global Hawk. The Global Hawk can travel 3,000
nm from its base to its operations area and stay aloft to loiter for 24 hours on
mission. That allows the Global Hawk to cover approximately 40,000 km2 of
ocean with its wide area SAR search capability [36]. These are certainly impres-
sive figures of merit. However, to put this in perspective in our context of mari-
time surveillance, consider that the area of the Pacific Ocean is over 150 million
km2 [37], a factor of more than 4,000 times the coverage of a Global Hawk.
The number of available platforms with this type of capability are limited, and
most ISR platforms have substantially less range, endurance, and search capabil-
ity, as we have discussed in earlier chapters. Even flawlessly executing the most
advanced strategies proposed in this book, it is still difficult to imagine how a
military force even as capable as the United States’ could maintain continuous
surveillance of an adversary’s Blue-water navy deployed on the open seas.
The only platforms capable of providing a vantage point from which to
conduct wide area surveillance of open oceans are space-based platforms that
rely on ship transponders that report ship locations for tracking. Compare the
ship-based AIS receiver range of 20 nm to the 3,630 nm wide swath that a satel-
lite in a 1,000-km orbit would have. It is estimated that such a satellite viewing
186 Naval ISR Fusion Principles, Operations, and Technologies

European waters would be able to track up to 6,200 AIS-equipped ships simul-

taneously [38]. The deployment of dedicated satellites capable of surveillance
of naval vessels and readily available for military tasking has historically been an
expensive endeavor (for example, Canada’s RadarSat constellation can provide
ship tracking, without the need for cooperative transponders). Furthermore,
until recently, the timeline for fielding such systems has been on average 7.5
years for a new system and 3 years for subsequent satellites of the same design
[39]. This has meant that space surveillance of the oceans has been largely lim-
ited to technologically advanced nations, primarily the United States and the
former Soviet Union with some limited capability by Europe and China [40].
The emergence of small-satellite technology over the last decade has completely
altered the calculus on space-based military surveillance of the oceans.
We have now (in the 2020s) entered an age where this kind of ubiquitous
sensing capability is available to any major nation state and indeed, even to
nonstate actors who are able to pay for commercial Earth observation services,
commercial services abound, providing not only imagery, but analysis as well.
One such vendor, Maxar, offers a huge range of imaging, sensing, and commu-
nication services, and is currently working with over 50 governments. Maxar’s
products include high resolution (less than 50 cm) visible imagery with less
than 5m geolocation accuracy from its WorldView satellites and high resolu-
tion and wide area radar imagery from the Canadian RADARSAT-2. Revisit
rates are at least once per day for the WorldView Legion imagery. Maxar also
provides complete analytics products. While much of this has to do with civil-
ian applications like change detection for agriculture or urban planning, and
disaster response, Maxar also advertises services for MDA and surveillance and
intelligence [41]. There is a potential emergence of commercial, space-based
services for maritime awareness using both imagery and RF signal mapping
(called RF imagery). This emergence can be expected to build on the history
of systems such as COSPAS/SARSAT which were first developed as coopera-
tive international space components that spawned commercial, end-user data
services. Even as early as the 1990s, private and pleasure craft could purchase
SARSAT beacons. Canadian RADARSAT services have been available for over
a decade to enable ship navigation in arctic conditions. Next, we will see near
real-time commercial maritime awareness MultiINT products using all avail-
able commercial sources.
The number of providers of such high-quality products continues to grow
rapidly with new products and services entering the market every month. For
instance, vendors such as LandInfo and Apollo Mapping resell WorldView im-
agery and provide their own analysis products. Commercial high-res imagery
can be purchased for as low as $10/km2, with priority tasking costing up to
$100/km2 [42]. This puts space-based reconnaissance in the reach of essentially
any nation state that wants it, whenever they want it, on an as-needed basis,
 Distributed Space Maritime Surveillance 187

without having to invest in a national space program. This is problematic for

larger nation states such as the United States who have been long used to hav-
ing information dominance regarding space-based intelligence. For example,
during the Afghanistan conflict in 2001 the United States was able to persuade
the French government to refuse to sell imagery from their SPOT satellites on
the open market where enemies could obtain it. The negotiation required the
Pentagon to purchase the imagery at a cost of $1.9M per month. In today’s en-
vironment of multiple international vendors and services, it is virtually impos-
sible to make such arrangements. In support of their missile attack on the U.S.
Al Assad airbase in Iraq, Iran purchased commercial satellite imagery to enable
targeting of personnel and aircraft at the base [43]. Such problems are causing
many to speculate as to whether it is justifiable to target commercial satellites
that are used for military purposes by an adversary. While the Law of War and
other legal arguments seem to justify attacking commercial assets used for mili-
tary purposes, the complexity of the multinational ecosystem of commercial
sensing may make such attacks infeasible from a diplomatic point of view [44].
China is emerging as a peer competitor to U.S.-based commercial vendors
with its Soar products [45] and has adopted a strategy of fielding smallsats for
surveillance of regions close to China. The Yaogan-30 constellation has put
nine groups of three satellites into LEO on a 35-degree inclination as of June
2021. The satellites will provide optical and radar imaging as well as signals
intelligence and may be similar in concept to HawkEye 360 [46, 47].

9.4 Surveillance Constellation Performance with Examples

ISR Constellations are designed to address intelligence, commercial, or global
situation monitoring and require a clear definition of the mission to establish
requirements to design the system. We consider in this section the scenario
previously developed in Chapters 2, 3, and 5 as a representative baseline chal-
lenge to develop an ISR constellation to illustrate the mission and system design
process (Figure 9.2).
The process proceeds in the seven general steps numbered in Figure 9.2
and described in the following paragraphs. Note that that the process as imple-
mented is iterative, with loop backs and refinements to develop alternatives and
a final recommended baseline solution. The following paragraphs describe the
seven steps:
Step 1: Mission goals are established to explicitly articulate the goals of the
fleet for the Kandigan security challenge—monitor (surveil) the activities in
the heavily contested Kandigan Straits area (and across a larger swath of con-
tested areas at similar longitudes). For this challenge problem, the constellation
must provide surveillance that meets the following derived objectives [O] that
188 Naval ISR Fusion Principles, Operations, and Technologies

Figure 9.2 A top-level constellation mission and system design process [48].

establish the effectiveness of the system to meet the mission of monitoring di-
verse sites and activities in the Kandigan Straits mainland (Red), island country
(Green), and contested island areas:

1. [O1]: At least daily monitoring of Kandigan naval activities. Prima-

ry—Red Nation ports. Secondary—Green Nation ports. Contested
islands.
a. Area of interest lies between 10S and 25N latitude—this affects
orbit design
2. [O2]: Provide strategic indications of changes:
a. Coastal edge changes, construction of airfields or port infrastruc-
ture: 4m to 8m EO or SAR.
3. [O1]: Surveil 24/7 for activity changes, surges:
a. A squadron of aircraft or major combatant vessel could arrive or
depart unobserved if the sampling interval is more than 4 hours.
4. [O2]: Surveil islands and surrounding naval/air activity and changes:
a. Ships: 10m to 20m resolution to detect, 3m to characterize type,
1m to identify specific ship;
b. Aircraft: 3m to characterize class, 1m to identify type.
5. [O3]: Maintain a baseline level of activities in the area to maintain
chain of custody on all naval maritime activities (up to 200 vessels of
interest) with maximum revisit of 3 hours, with an objective goal of
30 minutes; monitor background commercial maritime activity levels.
In addition, the system must:
 Distributed Space Maritime Surveillance 189

6. Maintain downlink reporting latency <10 minutes sensor collect-to-

report detection (ground stations or relay provided);
7. Provide spacecraft lifetime >2 years.

Step 2: Derive mission requirements. From the qualitative narrative mis-

sion goals, we must derive quantitative performance parameters. This process
proceeds from the objectives to derive the baseline observations requirements
in four general phases (Figure 9.3). Beginning with the three major objectives,
O1, O2, and O3, we specify the range of targets in the Kandigan Straits to
derive parameters that specify the dispersal of targets (phase 2) and then specify
the kinds of activities that must be monitored (phase 3). This allows us to derive
the quantitative surveillance parameters (phase 4) required of the constellations
system, and the key parameters (revisit rate, ground resolution, RF-persistence,
etc.).
Step 3 and Step 4: Design constellation space vehicles, links, and ground
enterprise. The constellation design uses a tool such as the commercial Ansys
Government Initiatives (AGI) Systems Tool Kit (STK) to develop orbital pa-
rameters and simulate the constellation operation to verify performance at the
specified target locations [49]. Additional analytic capabilities, as the Space
Mission Analysis and Design (SMAD) tool can be used to develop and analyze
spacecraft parameters. The flow in this section follows the general process used
in SMAD [50]. The analysis steps (Table 9.3) include the analysis to develop
the constellation size, mix, orbits, and type (e.g., Walker, Walker-D configura-
tions), the sensor requirements, and projected performance.

Figure 9.3 Derivation of ISR observation requirements from mission objectives O3 and O4.
190 Naval ISR Fusion Principles, Operations, and Technologies

Table 9.3
Summary of the Constellation and Space Vehicle Design Steps
Resulting
Analysis Steps Tradeoffs Considered Products
1. Select orbital ·· Apogee-Perigee, circular? ·· ISR coverage,
options for one ·· Inclination, orbit type, (e.g., SSO) in which the revisit, target
spacecraft satellite passes over any given point of the capacity.
planet’s surface at the same local time.
·· Revisit and coverage per spacecraft.
·· Trades between coverage, revisit, access.
·· Sensitivity to altitude, inclination.
2. Select constellation ·· Constellation revisit and coverage; ·· Constellation size,
type and size [STK] ·· Constellation type (e.g., Walker constellation configuration,
that consists of a set of satellites (t) that are and achieved
in circular orbits and have the same period and coverage.
inclination). Satellites (s) are evenly spaced in
each of the orbital planes (p) specified so that t
= sp.
3. Assess sensor ·· FOV access; ·· ISR sensor access
accesses to target for ·· Expected performance (GSD); performance.
sensor FOV ·· Sensor aperture vs performance.
4. Communication to ·· Downlink data volume from sensor for targets; ·· Communication
downlink data ·· Data rate and storage configuration;
·· Relay capability hypothesized; ·· Threat analysis.
·· Data link trades and impact on sensing capability;
·· Consider benefits of relay, cross-satellite links to
meet requirements.
5. Baseline ·· Orbital dynamics, geometry, maneuver needs ·· Spacecraft
analysis; velocity change
·· Estimate overall ∆V demands for lifetime. (∆V) demands.
6. ISR sensor ·· Assess altitude and elevation angles to ·· Optical design.
payload—optics investigate the impact of these decisions on the
design of the payload.
7. ISR sensor ·· Evaluate alternative sensor payloads, sensor ·· Sensor,
payload—physical pointing, slewing, and integration; processing, and
sizing and data rate ·· Analyze data rate, onboard store-forward, and downlink.
downlink requirements.
8. System sizing ·· Analyze, select type of attitude control; ·· Spacecraft
summary [SMAD] ·· Analyze, select primary power source; baseline design.
·· Analyze, select type of structure;
·· Analyze, select type of communications;
·· Analyze, select type of propulsion system
evaluate the baseline SC bus—preliminary sizing;
·· Assess deviations and compare to reference
example.
9. Communications ·· Evaluate mission data download strategy and ·· Downlink budget
analysis: effect on data rate, storage, power, and so on; and margin;
Downlink; ·· Assess downlink strategy impact on coverage and ·· Uplink budget and
Uplink. latency; margin.
·· Assess downlink and uplink parameters and link
margins.
10. Launch vehicle ·· Assess alternative launch options. ·· Launch method,
information schedule.
 Distributed Space Maritime Surveillance 191

Step 5 and Step 6: Perform constellation cost modeling. The spacecraft and
ground costs are estimated, as well as launch and operational costs to develop
a lifecycle cost. These estimates are compared to comparable prior systems and
established cost models. This step establishes the baseline design (constellation,
spacecraft, links, ground enterprise) and cost models.
Step 7: Effectiveness assessment. Throughout the design process, assess the
degree to which the entire system achieves the mission goals and refine the base-
line. While Figure 9.3 illustrates a single end-of-cycle assessment, the process
is performed throughout the design process to understand how design changes
(e.g., changes in spacecraft performance, number of spacecrafts, and ground
station coverage) impact the effectiveness to meet mission goals. This process
may include the use of simulations that represent the dynamic constellation
and its access to the Kandigan Straits and a simulation of dynamic naval forces
(of varying sizes and compositions in different scenarios) to determine the ISR
performance and contribution to fleet mission effectiveness.

9.5 Performance Study Results

With these tradeoff parameters in mind, let us consider a few potential designs
for this constellation. First recall our brief overview of the Flock and Capella
constellations in Sections 9.1.1 and 9.1.2 respectively. These offer examples of
existing smallsat constellations for visible imagery and SAR imagery respec-
tively. We will consider these as starting points for our baseline performance
and then suggest several variations on constellation configurations and modi-
fications of satellite designs to enhance key capabilities. Finally, we will look at
combining SAR and visible EO imagery in a MultiINT capability.
We ran simulations of each of the constellations below using the STK
modeling tool. We also considered slightly different designs for the CubeSats,
using a proprietary spreadsheet analysis tool developed at the Naval Postgradu-
ate School for doing first-order design analysis. We list the design points that
we considered, state the image resolution parameters and any comments on the
design, and provide a summary table (Table 9.4) of the revisit times.

9.5.1 Planet FLOCK–Based Solutions

A series of five alternatives with improving revisit rates were considered (Table
9.4) before arriving on the best solution (number 6). Note that options 5 and
6 included a modified Dove increased to 6U size and given off-nadir pointing
capability with a larger aperture. This capability enabled the reduction in num-
ber of spacecrafts in option 6.
192 Naval ISR Fusion Principles, Operations, and Technologies

Table 9.4
EO Constellation Options Considered
1. Planet’s initial Flock 3P 2. Big Flock—triple the 3. Modified Orbit Flock (88
(88 3U CubeSats): number of CubeSats in 3U CubeSats):
·· Spatial resolution of 3m; abovementioned example (264 ·· Same constellation of 88
·· Polar orbit, 97-deg 3U CubeSats): DOVE’s but incline orbit
inclination; ·· Orbit and resolution are to 23 degrees. Inclining
·· Nadir pointing only; unchanged; the orbit eliminates the
·· Revisit rate: 5.7 hrs. ·· Revisit rate improves to 2 wasted time that the
hrs. satellites are not focused
on our specific area.
·· Revisit time: 2.7 hours.
4. Modified Dove design (22 6U CubeSats):
·· Use a 6U CubeSat that improves on the 3U Dove in two important ways;
·· Larger aperture (90 mm) for sub 2m resolution imager;
·· Ability to point up to 30 degrees off-nadir axis;
·· Maintain 88 satellites in 23-degree orbit, revisit rate: 2.2 hours.
5. Best solution: optimal modified Dove (22 6U CubeSats):
·· Since the pointing capability did not substantially improve the revisit rate in the 88
satellite, inclined orbit constellation, we can deduce that we could reduce the number of
satellites and still meet our coverage goals; 22-satellite constellation was optimal;
·· Revisit rate 2.3 hrs.

9.5.2 Capella-Based SAR Solutions

A series of four alternatives were considered (Table 9.5), moving from a single
satellite to the full 36-satellite constellation, before arriving on a small 2-satellite
constellation in an inclined orbit that achieves the desired 2-hour revisit time.

9.5.3 Hybrid MultiINT Design

Finally, we will combine the best EO imaging constellation with the best SAR
constellation to provide both moderate resolution visible imagery with the day/
night, all-weather capability of SAR. This will provide a little over 2-hour revisit
rate with each sensing modality. Furthermore, if the orbits are timed properly
there will be a large portion of the time when there will be approximately one

Table 9.5
SAR Constellation Options Considered
1. Single Capella-like satellite 2. Capella 12 satellites: 3. Capella 36 satellites (Capella
in 97-deg SSO: ·· 1.3-hr revisit rate. full constellation):
·· 14-hr revisit time. ·· 0.8-hr revisit time.
4. Best solution: 2 satellites in 18-degree inclined orbit:
·· Same principle as inclining orbit of the Dove-like CubeSats, but the wider search swath allows for
a lower inclination;
·· Revisit rate: 2.6 hr.
 Distributed Space Maritime Surveillance 193

hour between one or the other sensing modalities. The interval between SAR
and EO/IR will not be constant because the constellations are in different or-
bital inclinations. Still, the gaps in coverage will be at most about 2 to 2.5 hours
and will average around 1 hour.
The advantage of the MultiINT constellation is that the wide area search
capability of the SAR can be used to identify potential targets of interest over
large areas of open ocean. In this mode the resolution is low (perhaps many 10s
of meters), but objects of interest thus detected can be handed off to the next
available EO imager. The EO imager, with its 1m to 2m resolution will have
a good chance of identifying a ship type. If cloud cover or darkness prevents
collecting EO imagery, the next SAR pass would be able to focus on a narrower
spot mode with higher resolution. Of course, to perform these handoffs prop-
erly requires a rapid transport layer to exchange cues, and an effective prediction
model for ship’s motion; we addressed such modeling in Chapter 8.
This example analysis did not address the performance of adding a sig-
nal capability such as HawkEye 360, but unless ships are operating completely
EMCON, then any RF emission (including navigation radar) can be used as an
initial list of locations for subsequent SAR or EO searches.

9.6 Conclusions
The prior sections provided a most basic introduction to the process of identify-
ing a maritime intelligence need, translating it to collection requirements, and
developing a suitable constellation of small satellites to perform the collection.
The analysis was intended to demonstrate the mission development process and
explain feasibility using commercial satellites. But it is important to consider
several factors in the process:

• We focused on the imaging satellites to assure coverage; we did not ex-

amine the use of the RF signal collection capability of Capella and its
contribution in detail. A follow-on analysis would estimate the coverage
and intercept contribution from such a collection capability [51].
• We did not examine the use of the relay constellation (introduced ear-
lier) to provide a transport layer to allow satellite cross-cueing. The com-
mercial Space-X Starlink constellation has demonstrated feasibility of
such capabilities and the DoD is developing a similar secure military
transport capability [52, 53].
• MultiINT analysis was not performed to determine the distributed
fusion capabilities, performance, and architecture, as we examined in
Chapters 6 and 7.
194 Naval ISR Fusion Principles, Operations, and Technologies

If we were constrained to only leverage an existing commercial constel-

lation, we can see that the fact that the commercial constellation may have
more general objectives, the constellation will not be efficient for a narrowly
constrained problem like ours. It also may not be possible to satisfy all the
performance desires since a commercial constellation has many customers to
satisfy besides naval forces. Of course, a dedicated maritime constellation can be
implemented by national organizations to support large fleets, or international
alliances. Finally, we need to be aware that the ubiquitous availability of com-
mercial satellite sensing products is a threat to naval forces who have enemies
that can leverage these commercial services to achieve parity in overhead ISR
capability. Yet, this was a most basic example, and new systems will be devel-
oped and deployed to provide the robust ocean surveillance and reconnaissance
capabilities envisioned earlier in this chapter.
Naval tactics must change dramatically to adapt to this new reality of
ubiquitous sensing. Major sea powers can no longer assume that they can op-
erate freely in Blue water and arrive at an adversary’s shores with the element
of surprise. Instead, they must assume that the location and status of all of
their ships could potentially be known by even the smallest of opponents. In
addition, the challenges of access to large amounts of intelligence data at sea
are being removed by satellite-based internet services. The playing field among
international navies will become substantially more level. Denial and deception
operations must take on a different character. Since denial of overhead intel-
ligence will be difficult, deception must take on greater importance. Navies will
need to consider how to position and posture forces in a way that creates some
ambiguity, forcing their adversaries to wonder about actual intent, even though
external observables can all be seen.
We have written about this elsewhere, making the following analogy be-
tween the need to search for adversary naval vessels and the game of Battleship
played on graph paper in the ready rooms of WWII carrier pilots:

An observable ocean will significantly change naval operations, and the

current game of ‘battleship’ will give way to a game of a chess—when all
can see the ‘gameboard’ clearly. Of course, if all can see the board (e.g.,
all have broadband, overhead surveillance and can see the opponent’s
fleet) the advantage will go to the strategist that can foresee future moves,
understand environmental factors, and make the most rapid, decisive
actions to achieve the advantageous position—consistently. In this future
environment, the victor must achieve a speed of decision and operational
advantage [54].
 Distributed Space Maritime Surveillance 195

Endnotes
[1] https://www.history.navy.mil/research/library/online-reading-room/title-list-
alphabetically/f/from-sea-stars.html#bookmark7.
[2] Bateman, A., “Technological Wonder and Strategic Vulnerability: Satellite Recon-
naissance and American National Security during the Cold War,” International Jour-
nal of Intelligence and CounterIntelligence, Vol. 33, No. 2, 2020, pp. 328–353, DOI:
10.1080/08850607.2019.1703926.
[3] https://www.history.navy.mil/research/library/online-reading-room/title-list-
alphabetically/f/from-sea-stars.html#bookmark7.
[4] Carlson, B., “NRO’s Historical, Current, and Potential Future Use of Small Satellites,”
National Reconnaissance Office, 2011.
[5] Table adapted from “Navy’s Needs in Space for Providing Future Capabilities,” National
Academy Press, 2005, Table 2.2, p. 42.
[6] Swartwout, M., “Reliving 24 Years in the Next 12 Minutes: A Statistical and Personal His-
tory of University-Class Satellites,” Proc. 2018 Small Satellite Conf., SSC18-WKVIII-03.
[7] Halt, T., “Smallsats by The Numbers: A Discussion of the Smallsat Industry,” Sat Maga-
zine, April 2019, http://www.satmagazine.com/story.php?number=1604295635.
[8] Commercial Orbital Transportation Services: A New Era in Spaceflight, NASA SP-2014-
617, 2014.
[9] Selva, D., and D. Krejci, “A Survey and Assessment of the Capabilities of CubeSats for
Earth observation,” Acta Astronautica, Vol. 74, 2012, pp. 50–68.
[10] Newman, J., C. Racoosin, W. Lan, R. Begley, N. Crawford, J. McGowan, and R.
McClenning, “GEOINT Small Satellite Constellation Study for Maritime Domain
Awareness,” Naval Postgraduate School, 2018.
[11] Van Ryswyk, M., “Planet’s New Rapid Revisit Platform to Capture Up to 12 Images
per Day,” Planet Pulse, June 9, 2020, https://www.planet.com/pulse/12x-rapid-revisit-
announcement/. See also Mason, J., “What Is Rapid Revisit and Why Does It Matter?,”
Planet Pulse, September 16, 2019, https://www.planet.com/pulse/what-is-rapid-revisit-
and-why-does-it-matter/.
[12] Wulder, M. A., et al., “The Global Landsat Archive: Status, Consolidation, and Direction,”
Remote Sensing of Environment, Vol. 185, 2016, pp. 271–283.
[13] https://www.usgs.gov/faqs/what-are-acquisition-schedules-landsat-satellites.
[14] Hallex, M. A., and T. S. Cottom, “Proliferated Commercial Satellite Constellations:
Implications for National Security,” Joint Forces Quarterly, JFQ 97, 2nd Quarter, 2020,
pp. 20–29.
[15] Erwin, S., “NRO Begins Journey Toward Commercial Data Program of Record,” Space
News, October 7, 2021. Using optical inter-satellite links, commercial imaging companies
could send data directly from their constellations to SDAs’ satellites.
[16] Erwin, S., “Maxar Still Confident Legion Constellation will be in Orbit in 2022,” Space
News, October 7, 2021.
196 Naval ISR Fusion Principles, Operations, and Technologies

[17] Planet Mission Statement; https://www.planet.com/

[18] Samuels, D., “Inside a Startup’s Plan to Turn a Swarm of DIY Satellites into an All-Seeing
Eye,” Wired, June 10, 2013, https://www.wired.com/2013/06/startup-skybox/.
[19] Planet Labs Investor Presentation, July 2021, https://www.planet.com/investors/
presentations/2021/investor-presentation-20210707.pdf.
[20] Federal Communications Commission, “Appendix A: Response to Question 7 of FCC Form
442 (Purpose of Experiment),” https://apps.fcc.gov/els/GetAtt.html?id=203932&x=.
[21] “Capella Space SAR System Performance,” Capella Space, September 11, 2020.
[22] “Capella Space SAR Imagery Products Guide,” Capella Space, September 11, 2020.
[23] Holmquist, R., “HawkEye 360: RF For Arctic Domain Awareness,” HawkEye 360 Inc.,
2021.
[24] https://www.he360.com/our-solutions/.
[25] https://en.wikipedia.org/wiki/Automatic_identification_system.
[26] Kurekin, A. A., B. R. Loveday, O. Clements, G. D. Quartly, P. I. Miller, G Wiafe, and K.
Adu Agyekum, “Operational Monitoring of Illegal Fishing in Ghana through Exploitation
of Satellite Earth Observation and AIS Data,” Remote Sensing, Vol. 11, No. 3, 2019,
p. 293.
[27] Longépé, N., et al., “Completing Fishing Monitoring with Spaceborne Vessel Detection
System (VDS) and Automatic Identification System (AIS) to Assess Illegal Fishing in
Indonesia,” Marine Pollution Bulletin, 2018, Vol. 131, pp. 33–39.
[28] https://www.newspace.im/constellations/hawkeye-360.
[29] CaJacob, D., N. McCarthy, T. O’Shea, and R. McGwier, “Geolocation of RF emitters
with a formation-flying cluster of three microsatellites,” HawkEye 360 Inc., 2016.
[30] Erwin, S., “DoD Space Agency to Create Marketplace for Commercial Satellite Data,”
Space News, June 22, 2021.
[31] “National Defense Space Architecture (NDSA) Systems, Technologies, and Emerging
Capabilities (STEC)” Broad Agency Announcement (BAA), SDA, HQ085020S0001,
January 21, 2020, https://sam.gov/opp/89fc1a2de9314f98b2b3f63ee3b4600c/view.
[32] “Space Transport Layer Tranche 0 Statement of Work (SOW),” SDA, HQ085020R0001
Attachment 1 SOW, March 26, 2020. See also National Defense Space Architecture
Tranche 1 Demonstration and Experimentation System (T1DES) Program Solicitation,
June 21, 2022; https://sam.gov/opp/567d447a0f024c6b8aea0a7b1a4707e2/view.
[33] Space Development Agency Next-Generation Space Architecture, Request for Information,
SDA-SN-19-0001, July 2019, p. 2.
[34] On the tasking of commercial imaging spacecraft, see Werner, D., “NRO Shares Plans
for Commercial Imagery Acquisition,” Space News, June 4, 2019, https://spacenews.com/
nro-shares-plans-for-commercial-imagery-acquisition/.
[35] HQ085020R0001, “Space Transport Layer Tranche 0,” Space Development Agency, SOW,
HQ085020R0001, March 26, 2020, https://imlive.s3.amazonaws.com/Federal%20
Government/ID30063299141661607590122820076958924261/Attachment%20
 Distributed Space Maritime Surveillance 197

1%20-%20SDA%20Transport%20Statement%20of%20Work%20%28DRAFT%29.
pdf.
[36] https://irp.fas.org/program/collect/global_hawk.htm.
[37] https://en.wikipedia.org/wiki/Pacific_Ocean.
[38] Eriksen, T., G. Høye, B. Narheim, and B. J. Meland, “Maritime Traffic Monitoring Using
a Space-Based AIS Receiver,” Acta Astronautica, Vol. 58, No. 10, 2006, pp. 537–549.
[39] Davis, L. A., and L. Filip, “How Long Does It Take to Develop and Launch Government
Satellite Systems,” International Cost Estimating and Analysis Association, AEROSPACE
REPORT, 2015.
[40] https://cimsec.org/importance-of-space-in-maritime-security/.
[41] https://www.maxar.com/.
[42] https://apollomapping.com/imagery-dem-price-lists.
[43] https://www.cbsnews.com/news/iran-missle-strike-al-asad-airbase-60-
minutes-2021-02-28/.
[44] Dunlap, C., J. D., “Are Commercial Satellites Used for Intelligence-Gathering in Attack
Planning Targetable?” March 5, 2021, https://sites.duke.edu/lawfire/2021/03/05/are-
commercial-satellites-used-for-intelligence-gathering-in-attack-planning-targetabl.e/.
[45] https://www.aaas.org/resources/high-resolution-satellite-imagery-ordering-and-analysis
-handbook#VI.%20Image%20Ordering.
[46] Jones, A., “China Launches New Group of Classified Yaogan-30 Satellites” Space News,
June 18, 2021, https://spacenews.com/china-launches-new-group-of-classified-yaogan-
30-satellites/.
[47] https://en.wikipedia.org/wiki/Yaogan.
[48] For a more detailed example, see Ravanbakhsh, A., and S. Franchini, “System Engineering
Approach to Initial Design of LEO Remote Sensing Missions,” 2013 6th International
Conference on Recent Advances in Space Technologies (RAST), 2013, pp. 659–664, doi:
10.1109/RAST.2013.6581292.
[49] See the AGI description of this tool at https://www.agi.com/products/stk.
[50] Space Mission Engineering: The New SMAD, J. R. Wertz and W. J. Larson (eds.), Space
Technology Library, Vol. 28, 2011.
[51] For a similar analysis with HawkEye 360 see the presentation, Lowe, J., STK
Multi-Int Mission Planning, AGI, June 11, 2019, https://www.youtube.com/
watch?v=zhSGLNVyhVA.
[52] Erwin, S., “U.S. General: Starlink in Ukraine Showing What Megaconstellations Can
Do,” Space News, March 8, 2022, https://spacenews.com/u-s-general-starlink-in-ukraine-
showing-us-what-megaconstellations-can-do/.
[53] Erwin, S., “Lockheed Martin, Northrop Grumman, York Space Selected to Build DoD’s
Internet-in-Space Constellation,” Space News, February 28, 2022, https://spacenews.com/
lockheed-martin-northrop-grumman-york-space-selected-to-build-dods-internet-in-
space-constellation/.
198 Naval ISR Fusion Principles, Operations, and Technologies

[54] Williamson, W., III, “From Battleship to Chess,” Naval Institute Proceedings, Vol. 146,
No. 7, July 2020, p. 1409
10
Future Technologies to Enable All-
Domain
Naval officers have always been familiar with the concept of the watch—a pe-
riod during which a sailor is assigned a role to perform a duty or keep a watch
to persistently oversee an assignment. Of course, the traditional watchkeeper
assignment was what today we call ISR—to maintain constant surveillance of
the ocean’s horizon.
Today, we use the term technology watch to describe surveillance of sci-
ence and the resulting technologies that may disrupt naval strategy, operations,
and tactics. This chapter introduces the critical role of a naval ISR technology
watch, and example technologies that are under surveillance and being moni-
tored by major navies. Technology is monitored, for example, by organizations
such as the U.S. Office of Naval Research, the Australian DST-G Maritime
Division, and China’s Naval Research Academy [1].
But the concept of a technology watch is not new. It was R.V. Jones, the
head of British Scientific Intelligence in World War II, and recognized devel-
oper of methods of empirical technical intelligence, that articulated the concept
of watch. The watch required monitoring emerging technology and the basic
steps required by an adversary to introduce new technologies in warfare [2]:

1. General scientific research of an academic or commercial nature oc-

curs, which causes:
2. Someone in close touch with a fighting service, and who is aware of
service requirements, to think of and promote an application of the
results of academic research. If this application is considered promis-
ing, then:
199
200 Naval ISR Fusion Principles, Operations, and Technologies

3. Ad hoc research and small-scale trials are performed in a service labo-

ratory. If these are successful, then:
4. Large-scale service trials are undertaken, which may lead to:
5. Adoption by the service.

Though writing about the process applied by the British and Americans
from the 1940s and 1950s, the process today adopts the following essential core
elements. Scientist and technologists perform analyses that include:

• Science research tracking to understand emerging science (physical sci-

ences, human behavior, cognitive science, and others);
• Naval application analysis to identify the ways breakthrough capabilities
can be applied to naval operations, and determine how they will disrupt
tactics, operations, and strategy;
• Technology tracking to monitor the state-of-the-art and an adversary’s
progress.

A quick review of naval operations since World War II demonstrates the

shifting emphasis on naval operations and the significant disruptive impact of
new technologies that brought us to the twenty-first century (Figure 10.1).
Now, accelerating technologies are again disrupting naval operations because
of increased ocean transparency (ISR), reduced decision timeframes (C2), and
fast weapons (lethality). In addition, the domains of ISR and C2 are expanding,
requiring the integration and coordination of warfare across many domains.
Writing on the effect of future technology on naval operations, Vice
Admiral Arthur Cebrowski (USN) pointed out that the ultimate importance

Figure 10.1 Disrupting technologies and strategic objectives of naval operations over time.
 Future Technologies to Enable All-Domain 201

of commanders’ accurate perception, correct decision making, and speed of

execution:

Our present information age illustrates the close coupling between tactics
and technology and portends a sharply increased power of tactics in
effecting operations and strategy. While strategy may seek to control the
scope, pace, and intensity of a conflict, tactics controls the very powerful
second derivative, that is, the rates of change that affect men’s minds where
wars are won and lost [3].

The following sections in this chapter describe the process used to scan
the (time) horizon for technologies that will change, disrupt, or otherwise im-
pact naval ISR and operations in the future (Section 10.1) and two specific
examples of emerging technologies that will certainly disrupt ISR in the near-
future (Sections 10.2–10.3).

10.1 Naval Technology Scanning to Avoid Surprise

New technologies have the potential to disrupt naval strategy, operations, and
tactic, so we use the term horizon scanning to scan the fields of science to iden-
tify the new or emergent scientific concepts and discoveries that may lead to en-
abling or disruptive technologies—those that enable new naval capabilities and
disrupt current operations. Once identified, we maintain a technology watch to
characterize the technologies in a known field, and their potential applications
in naval warfare. Awareness of the continually emerging science and technology
(S&T) fields are essential to avoid technology surprise—the sudden advance and
application of an applied science or technology, which for some period provides
an adversary with an advantage [4]. Key technologies being sought by navies are
enabling the following operational advantages:

• Broader scope and persistence in surveillance (all-domain sensing, mas-

sive computing, storage, sensor, and command networking);
• Shortened attack timelines (shortened hypersonic weapon transit times,
directed energy weapons, and rapid autonomous detection-track);
• Longer range engagement (long-range engagement enabled by autono-
mous unmanned platforms and networked autonomous sensing).

The defense warning community has identified four categories of surprise

mechanisms that all apply to the naval technology watch process (Table 10.1).
Key terms associated with the scanning and watch process that address
more general and broad activities within defense and naval operations include:
202 Naval ISR Fusion Principles, Operations, and Technologies

Table 10.1
Defense Categories of Technology Surprise
Type 1: A major technological Example: Introduction of the atomic bomb at the end of
breakthrough in science or World War II. Introduction of Earth satellites for national
engineering. reconnaissance in the U.S. Corona program. Introduction
of stealth technology to defeat radar detection. (These are
generally rare events, enabled by experts within the field.)
Type 2: A revelation of secret Example: The successful testing of hypersonic missiles by
progress by a second party which Russia and China has changed the threat balance. These
may have an unanticipated impact. missiles, launched from naval vessels will enable prompt,
long-range engagement and change the balance of naval
fleets.
Type 3: Temporal surprise, Example: Progress in foreign nuclear weapon, ballistic, and
when a party makes more rapid hypersonic missile programs where surprise is facilitated
development or advancement by technology transfer that accelerates progress beyond a
in a particular technology than traditional linear development cycle.
anticipated.
Type 4: Innovative technology Example: Introduction of unmanned surface vessels and
application. Such innovations do underwater vehicles into naval operations to perform
not necessarily require technical operations deemed too risky for manned vehicles.
expertise, but rather the creativity
to use available resources in a new
way.
From: [5].

• Technology forecast: Systematic methods to anticipate and understand

the potential innovation, discovery, characteristics (direction, rate, ma-
turity, readiness levels), and effects of technological change to areas of
adoption and application.
• Technology watch and horizon scanning: Technology watch is the charac-
terization of activity in a known field, and horizon scanning focuses on
identifying new or emergent concepts.
• Defense warning: The process of identifying sources of increasing threats
to naval interests (including science and the application in technology)
and identify opportunities to affect adversary behavior prior to and in
the early stages of a crisis.

Many methods have been developed to perform these functions because of

the board concerns about the offset and threat posed by disruptive technologies
[6].We describe here one detailed process developed by RAND: the STREAM
process (systematic technology reconnaissance, evaluation, and adoption meth-
odology) [7]. The process (Figure 10.2) compares emerging technologies to
alternatives on the basis of their likely effects on naval mission goals, including
consideration of barriers to implementation.
 Future Technologies to Enable All-Domain 203

Figure 10.2 The five steps of the STREAM process applied to naval technologies [8].

STREAM was designed to help agencies:

• Assess current and potential technologies according to characteristics

directly relevant to agency missions and to the policy environment in
which agencies operate;
• Incorporate such assessments more effectively into the existing agency
functions, including planning, system maintenance, and operation;
• Better account for the uncertainties inherent in the distribution, adop-
tion, implementation, and operation of proven technologies as well as
prospective future technologies.

The process proceeds in five steps from framing the problem to the iden-
tification of technologies that emerges from science and applying it to the
agency’s needs. Next, the technologies are characterized, defining the drivers,
204 Naval ISR Fusion Principles, Operations, and Technologies

risks, and barriers to adoption and costs. The tradeoffs between alternative ap-
proaches are conducted and reported to decision makers to select actions (e.g.,
commitment of resources to development of new capabilities). The process also
provides a basis to conduct a technology watch to observe adversary technolo-
gies with disruptive naval implications.
Following the scanning process, an assessment is performed to prepare
a strategy to deal with the emerging technology capability to avoid surprise
deployment. A 2008 Defense Science Board study identified five steps that,
integrated together, constitute a robust approach to managing surprise [9]:

• A scanning and sifting process that narrows the many possibilities to the
most worrisome few;
• A Red capability projection function that takes a deeper dive on the wor-
risome few through analysis, simulation, experimentation, and/or pro-
totyping;
• A net assessment process in which the deeper understanding of Red,
gained through capability projection, is played against Blue capabilities
in order to assess the degree to which the nation can address the threat
or adapt capabilities already in hand;
• An options analysis team to provide an unbiased evaluation—or “rack and
stack”—of the alternatives should Blue capabilities prove inadequate;
• An ability to produce a decision package that can be acted upon by senior
leadership.

In this book, our focus on naval maritime ISR is a subset of the wide range
of technologies surveilled by naval researchers [10]. Table 10.2 provides a cat-
egorized list of important technologies that navies have placed under watch in
the area of ISR impact because of their potentially disruptive effect. Of course,
these are also the areas of R&D investment and form the basis for research strat-
egies, development planning, and R&D portfolio management [11].
In the following sections we highlight just two of many critical technol-
ogy areas that will disrupt current naval operations when the technologies can
be scaled up, operationalized, and deployed. We have chosen these two, one
near-term and one long-term, to highlight the diversity and breadth of tech-
nologies that navies must consider. In addition to these two, we have previ-
ously introduced advanced wireless technologies (Section 7.2), AI and machine
learning technologies (Chapter 8), and small satellite constellation technologies
(Chapter 9).
 Future Technologies to Enable All-Domain 205

Table 10.2
Basic List of Naval ISR Technology Watch Categories
ISR Area Technology Application Areas
ISR platforms Undersea sensing (EM, ·· Seabed, undersea surveillance nets;
acoustic, energy) and ·· Undersea-to-space communication.
communications
Autonomous unmanned ·· Close-in surveillance, sensor deployment, and
vehicles OTH relay.
Large satellite constellations ·· Persistent maritime surface vessel tracking.
Distributed ISR EO/IR, radar, MASINT ·· Distributed network sensing across traditional
sensors sensing manned and unmanned-autonomous
platforms. Enhanced target recognition and
situation assessment.
Acoustic, magnetic sensing
Quantum sensing
Distributed AI ·· Automated and independent sensing, fusion,
ISR networks, and reporting.
processing Secure, wireless advanced ·· Sensor networking above the ocean surface
generation communication and undersea wireless (acoustic, optical, RF)
comm nets.
Advanced cryptography, ·· Improved communication channel security:
blockchain custody confidentiality, integrity, and assurance.
High-performance, quantum ·· Advanced modeling and simulation;
computing ·· Cryptographic and optimal search solutions.
ISR human- Augmented reality ·· Operator immersion in high dimensional
machine information.
analysis-C2 Human-machine teaming ·· Optimal coordination of human and machine
learning, reasoning, and problem solving in
complex contexts.
Weapons for ISR Hypersonic aircraft, ·· Reduced time-to-target engagement and
targets missiles improved OTH selection of targets and
endgame guidance.
Autonomous target
identification, selection,
and aiming

10.2 Undersea and Seabed Surveillance

While large constellations of networked satellites will render the ocean surface
very transparent, navies are also seeking technologies to enable persistent sur-
veillance of the undersea and seabed. These technologies would significantly
disrupt submarine and seabed warfare, providing an advantage to the navy that
has greater ISR visibility under the sea [12].
Traditional means of penetrating the deep included seabed sonar nets like
the U.S. Navy SOSUS which uses underwater hydrophone arrays to detect and
track submarines over large areas. The newer U.S. Navy ADS is deployable by
ship or submarine in deep or shallow water. The ADS underwater sensors link
206 Naval ISR Fusion Principles, Operations, and Technologies

their data by cable to surface processing-communications buoys. Once targets

are located within an area, submarine targets can be localized by air-dropped so-
nobuoy grids reporting to overhead patrol aircraft, or by ships deploying towed
array sonar systems. Because the main targets of attack submarines have been
strike groups and convoys of ships, the primary ASW sensor platforms have
included smaller ships (e.g., frigates) with towed arrays, attack submarines with
acoustic sensors, and airborne platforms with dipped sonars or sonobuoys to
extend the sensor barrier around high value units (HVUs). These systems are
limited in coverage, persistence and tracking performance in Blue water oceans,
and large contested seas that have dense commercial maritime traffic volume.
DMO seeks to distribute these assets, expand the detection range around a
HVU, and reduce risk using USVs.
The ISR missions for unmanned systems include all forms of surface and
undersea warfare: ASW, antiseabed warfare (countering operations to, from,
and across the ocean floor, generally attacking/defending seabed infrastructure,
such as undersea cables, oil heads, and pipelines), and counter UUVs [13]. Of
course, the close-to-surface systems can also monitor surface maritime activity,
surface environmental state, weather, and even detect aircraft overflights (Table
10.3).
The types of unmanned or unattended sensor platforms that seek ad-
vanced technologies to increase endurance, persistence, and sensor coverage in-
clude several categories depicted in Figure 10.3 and described as follows:

Table 10.3
ISR Missions Supported by Unmanned Platforms
ISR-Related Mission Functions
Extended range surveillance Deploy or provide platform for sensors to extend surface,
subsurface sensing with relay to the fleet via satellite. Perform
endurance scouting and surveillance of high-risk transit routes.
Surveillance sensor deployment Deploy leave-behind acoustic arrays has been demonstrated by
the ADS, which uses autonomous underwater vehicles (AUVs) to
deploy its sensor arrays.
Near-land, harbor-monitoring UUVs deploy sensor packages at near-land locations: outside
ports, shipping channels, and naval installations. Deploy
protective warning sensors for special force operations.
Seabed warfare Monitor undersea infrastructure and vessels that threaten nodes
and links; provide targeting for adversary vessels.
ASW and antisurface warfare Apply USVs and UUVs with sensors to detect, cue other assets,
(ASUW) localize targets, and reduce detection-to-engagement time in
both deep and shallow waters.
Mine warfare; mine Denied area MCM operations deploy UUVs from nuclear attack
countermeasures (MCM) submarines (SSNs) within the denied areas or by launching
longer-endurance UUVs from surface ships.
From: [14].
 Future Technologies to Enable All-Domain 207

Figure 10.3 The surface and undersea ISR sensor platforms.

1. Large seabed sensor networks. Large passive underwater sonar arrays

were first developed and refined throughout the 1960s to 2000 to
detect and track submarines detecting the low frequency acoustic and
electromagnetic signals they emitted. Linear arrays of acoustic sensors
(hydrophones) were placed across the ocean floor at strategic locations
(gaps that enable submarine passage) with undersea cables leading to
shore facilities that monitored individual hydrophones or steerable
beams formed by signal processing for detection and tracking [15].
More advanced systems integrate fixed and mobile (ship towed linear
array) systems that terminate at shore facilities. Advanced signal pro-
cessing enables search for signatures across a multidimensional space
of arrays, bearings, frequencies, and time [16].
2. Autonomous seabed networks. Networks deployed in denied littorals or
deep Blue water include power and control units with acoustic sensing
and ACOMM to intercommunicate for distributed detection and to
relay detections to UUVs above or to surface buoys that relay detec-
tions and tracks to satellite networks. The seabed network’s autonomy
monitors the environment, conserves power, performs distributed de-
208 Naval ISR Fusion Principles, Operations, and Technologies

tection and tracking, relays messages to other platforms, and reduces

emissions when threatened.
3. UUVs. UUVs provide a platform to enable autonomous extended
range ISR both underwater (acoustic and electromagnetic EM sens-
ing) and when surfaced (signals intelligence, electronic intelligence,
and imagery collection). Small and medium sized UUVs provide ISR
at long ranges to surveil surface ships or submarines, while swarms of
UUVs could perform a standoff counter-ISR mission against these
targets to spoof signatures or deceive navigation systems. Extra-large
UUVs (XLUUV) hold the promise of deploying seabed sensor net-
works and carrying ASW surveillance missions currently assigned
to manned submarines because of their large payload bays that can
accommodate sensors (e.g., side-scan sonar, towed array), and com-
munication payloads (to communicate with seabed sensors, surface
vessels, and space networks).
4. Submarines. It is important to recognize the presence of manned sub-
marines in their critical ISR role to employ acoustic, and special sen-
sors to detect and track submarines, or surface sensors (optical, radar)
to locate surface vessels. Added to the underwater communication
net, they may also contribute their organic sensor data to the network.
5. USV. These semiautonomous platforms include modular sensor ships
(medium USV) ranging up to larger antisurface warfare ship (large
USV). All contribute to long-range ISR for surface vessels, subma-
rines, and aircraft using electromagnetic, radar, and acoustic sensors.
The vessels can also provide terminal relay for long-range coopera-
tive engagement, relaying to a missile the terminal coordinates of a
maneuvering target obtained by other sensors (e.g., spaceborne SAR).
Shorter range USVs can support shorter range ISR missions in littoral
waters.
6. Unmanned surface and subsurface sensor nets. Autonomous networks of
floating buoys on the surface of the water can perform several func-
tions: (1) sense the environment and activity on the air, surface, or
subsurface around the buoy, (2) receive information from undersea
sensors, (3) act as a gateway to relay information to other subsurface
vessels, or to seabed sensors, and (4) act as a gateway to pass informa-
tion on to space constellations for further relay to fusion nodes. Large
networks have been deployed and these naval applications require
year-long operational life, low cost, and multiple sensing capabilities
to be viable. The DARPA Ocean of Things project, for example, seeks
to deploy up to 15,000 free floating, low-cost buoys to perform wide-
scale ocean environmental sensing and maritime surveillance [17]. It
 Future Technologies to Enable All-Domain 209

is worthy to note that maritime ISR may also benefit from enhanced
environment sensing to support naval operations. Information that
characterizes such factors as temperature, depth (bathymetry), seabed
clutter, and more, may impact naval operations including the detec-
tion of targets. Advances in technology such as acoustic tomography
may impact the concept of DMO promoted in this book.
7. Unmanned surface gliders. Surface UUVs that capture the kinetic en-
ergy of wave motion to create electrical power provide the benefits of
endurance (although they may require periodic refueling), and of the
ability of the station keep, and even provide limited maneuverability.
Gliders have been applied to oceanographic sensing; these platforms
can also provide tactical ISR surveillance of littoral and far-forward,
areas. The application to ISR has long been studied to provide surveil-
lance of far-forward, littoral areas, using gliders to achieve persistence,
stealth, and OTH launch with rapid transit to distant operating area
[18].

New enabling technologies are being developed to increase the cover-

age volume, detection performance, and persistence of ISR systems that will
cover the depths from seabed to the surface. To accomplish this operational
disruption, technologies must address several significant technical challenges to
achieve new capabilities.
Undersea communications networking. Communication of sensor data
from undersea sensors to other undersea platforms (for relay) or to surface
buoys or vessels remains a challenge to implement large, distributed undersea
sensor networks. For example, a U.S. Department of the Navy and U.K. Royal
Navy partnership for innovation defined the operational need for a high band-
width long-range underwater cooperative network as:

Cooperative communications systems that provide the digital backbone

underwater for data transfer, mission monitoring and mission management
between AUVs, in-water deployed networks, nodes or buoys. Network
may use RF gateways, acoustic, non-acoustic or other technologies and
achieve long-range application-specific connectivity ranging from 500m
up to 1000km. Solutions will need to interface with, (or exploit) above
water networks to form end-to-end solution [to] achieve high bandwidth
speeds of at least 1 kbps going to 100 bps at extended ranges [19].

Several alternative communication channels exist and are described in

the following paragraphs. Figure 10.4 distinguishes the typical range-data rate
operating envelopes of each alternative (optical, acoustic RF) as limited by
physical constraints; the figure does not imply that current operational links
210 Naval ISR Fusion Principles, Operations, and Technologies

Figure 10.4 Underwater communication performance [20].

are employed in region. The circle in Figure 10.4 shows the desired operating
range (0.1–1 Mbps and 1–10 km) for many naval applications to communicate
between buoys, UUVs, submarines, and seabed networks. The following lists
summarizes each of the three transmission methods in the figure:

• ACOMM: Modulated acoustic signals have demonstrated viable under-

water communications at low data rates for tens of nautical mile ranges.
Acoustic modems (modulator-demodulator transceivers) can be confor-
mally mounted on UUVs, hung on cables from surface vessels or buoys
that act as gateways from the network to satellites, or floated above sea-
bed sensors on cables to create a network of ACOMM transceivers to
intercommunicate. The challenges to go beyond state-of-the-art point-
to-point communication to distributed, multiuser networks must ad-
dress several issues: (1) range—ranges well beyond 40 nm are desired to
enable large nets, (2) covertness—low probability of intercept must ap-
ply forms of spectrum spreading or beam-pointing, both challenges for
the low bandwidth of acoustics, and (3) reliability—methods of channel
coding are required to achieve reliable digital communication that maxi-
mizes the use of the low channel bandwidth. ACOMM standards have
been developed to provide protocols for secure communications and IFF
between undersea vehicles as undersea traffic increases [21].
• RF: RF communications underwater are extremely limited due to the
significant attenuation of electromagnetic waves in saltwater (high elec-
trical conductivity) that limits the propagation range to a few kilometers
or less in seawater [22]. The severe limitations on data rates and propa-
 Future Technologies to Enable All-Domain 211

gation distance simply due to physical properties reduces the promise

of research breakthroughs to apply RF links for all but short distances.
• Optical communications: Commercial wireless optical communica-
tion systems can provide data rates on the order of 10 Mbps at ranges
<200m, and significant research has been conducted to achieve higher
data rates (up to 100 Mpbs) at longer ranges using blue-green lasers to
enable aircraft lasers to penetrate the air-water inface to communicate
with submerged submarines. This remains an area of research to achieve
high data rate results at longer undersea ranges.

In addition, we also consider two other transmission options:

• Wired cables: Fiber-optic cables (or tethers) are employed for towed sen-
sor systems, with cables as long as 1 to 5 km. But the use of these be-
tween undersea sensors and buoys, or even between UAVs and ships
or buoys to enable wideband communications have the obvious disad-
vantage for mobile sensors of physical connections, tangling risk, and
line-length. For this reason, mobile systems seek wireless (acoustic, RF,
or optical) solutions.
• Expendable buoys: For one-time seabed-to-surface (or submarine-to-
surface) store-and-forward communications, expendable buoys may be
floated to the surface to forward messages to satellites for relay to data
fusion nodes. Of course, the magazine of expendable buoys stored at the
seabed or in a UUV or submarine limits this mode of communication
to only the most critical messages. Such a solution is an engineering
solution rather than a new technology, but self-deploying smart buoys
could provide one engineering approach to realize a gateway interface
for seabed to space comms.

Underwater navigation: Underwater navigation in UUVs and submarines

is performed by inertial measurement units (IMU) using ring laser gyros
and accelerometers, updated by star tracking, GPS, or undersea ranging
networks. The continued refinement of these capabilities to extend range
of performance is desired for long-range UUVs. Relative acoustic posi-
tioning between UUVs has been studies to allow formations of UUVs to
maintain separation while maneuvering [23].
Platform autonomy: Autonomous operations require all platforms (buoys,
gliders, USVs, UUVs, and seabed networks) to (1) manage energy, (2)
monitor for threats and manage comm emissions to avoid detection, (3)
manage sensors, (4) fuse information from other sensors and relay new
212 Naval ISR Fusion Principles, Operations, and Technologies

detect/ID information across the network, (5) maneuver or station keep

as required, and (6) manage communications on the net and spectrum
utilization. These capabilities will apply advanced reasoning and learning
technologies to adapt to the environment (e.g., natural and manmade
threats) and optimize the use of organic and networked sensors to con-
tribute to fleet ISR.
Platform power and propulsion: Research in more efficient onboard bat-
teries, fuel cells, and other energy sources (wave, solar, etc.) will enable
longer ranges. Improved station keeping and maneuver of all unmanned
vehicles.
Operational considerations: The use of UUVs for sensor deployment and
sensing is not entirely covert, and the very use of some of these systems
may provide a cue to adversaries and even be considered threatening. A
study by the National Academy of Sciences noted:

Rules of engagement. How is it determined when to deploy a distributed

system during pre-hostilities in a potential battlespace? Naturally this
involves the complexity and cost of the system. Some operations—such
as activating a barrier—could be considered hostile. How much tactical
advantage is an adversary allowed to accrue before U.S. forces fire the first
shot to prevent U.S. ships and sailors from being at unacceptable risk?
Guidelines need to be developed for employing these systems, which in the
aggregate will be expensive, but they certainly should be used if U.S. ships
and sailors will be at risk [24].

One operational concept for these technologies foresees the following

stage of fleet operations in a denied area operation [25]:

1. Environmental survey is performed by deployed drifters (surface-ve-

locity profile) and other UUV and surface sensors to perform acoustic
assessments, bathymetry, and sea floor topology, to estimate required
sensor mix of fixed and mobile sensors.
2. Fixed and mobile sensor nodes are launched from submarines, littoral
combat ships, or USVs to provide a network with optimum surveil-
lance coverage. The UUVs enter semidormant state as temporarily
fixed or drifting nodes, awaiting guidance.
3. Mobile sensors nodes are commanded to maneuver as the tactical or
environmental situation changes to maintain surveillance coverage.
4. Target initial detections are communicated to the network via the net-
work (acoustic or optical undersea links) to buoyant gateways that
relay via RF link to an overhead constellation of relay nodes.
 Future Technologies to Enable All-Domain 213

5. The mobile asset wolfpack responds to the detection to track and

achieve weapon firing criteria.

10.3 Quantum Technologies in Naval ISR

Naval ISR inherently relies on remote sensing, computational power, and se-
cure communications to see further, deeper, and communicate instantaneously
over dynamic communication networks. The power of computers has grown
at an amazing pace since the invention of the transistor and has been doubling
in power about once every 2 years following Moore’s law [26]. It is important
to note that the computer performance is improved by reducing the size of the
electronics in the computer processors. Unfortunately, this reduction cannot
continue forever because as the size of electronic components starts to approach
the atomic scale, quantum effects hinder the performance of computers, and
hence set a limit on the computing technology. As the demand for faster com-
puter continues, a radical new approach is required.
A possible framework is one that performs computation using quantum
theory instead of classical mechanics. In the early 1980s, the Nobel prize-win-
ning physicist Richard Feynman said “…Nature isn’t classical, dammit, and if
you want to make a simulation of nature, you’d better make it quantum me-
chanical, and by golly it’s a wonderful problem, because it doesn’t look so easy”
[27]. Feynman observed that classical computers are inherently incapable of
simulating the evolution a quantum system due to the large number of variables
required to completely represent such a system. No one then knew how to build
a quantum computer nor how to use quantum mechanical effects to reveal the
power of quantum computation. However, in 1994, Peter Shor surprised the
community with a novel algorithm capable of exploiting quantum mechanical
effects to find the prime facto­­rs of an integer in a time that scales as a polyno-
mial with the size of the input, in contrast to the exponential time taken by
classical algorithms [28]. As a result, the security of the current cryptosystems,
which depends on the difficulty of factoring prime numbers, became highly
questionable. Shor’s algorithm sparked a great deal of interest in physicists, en-
gineers, computer scientists, and mathematicians throughout the world to col-
laborate with a common goal of developing a computer that uses the laws of
quantum theory. The next paragraphs describe the areas that have the potential
of significantly impacting ISR by introducing revolutionary new capabilities
and performance in sensing, communication, and computation.
A recent study by the National Academy of Sciences concluded that this
technology is not imminent, but will be significantly disruptive across all na-
tional security domains, including the naval DMO approaches espoused in this
book:
214 Naval ISR Fusion Principles, Operations, and Technologies

Quantum computing research has clear implications for national security.

Even if the probability of creating a working quantum computer was low,
given the interest and progress in this area, it seems likely this technology
will be developed further by some nation-states. Thus, all nations must plan
for a future of increased QC capability. The threat to current asymmetric
cryptography is obvious and is driving efforts toward transitioning to
post-quantum cryptography… Any entity in possession of a large-scale,
practical quantum computer could break today’s asymmetric cryptosystems
and obtain a significant signals intelligence advantage. While deploying
post-quantum cryptography in government and civilian systems may help
protect subsequent communications, it will not protect communications
or data that have already been intercepted or exfiltrated by an adversary.
Access to prequantum encrypted data in the post-quantum world could
be of significant benefit to intelligence operations, although its value
would very likely decrease as the time horizon to building a large-scale
QC increase. Furthermore, new quantum algorithms or implementations
could lead to new cryptanalytic techniques; as with cybersecurity in general,
postquantum resilience will require ongoing security research [29].

10.3.1 Quantum Computing

A classical computer is based on bits that can take on only two values, 0 and 1.
Bits are like the status of a light bulb controlled by a switch: the bulb is either
off or on. There are no states in between. In addition, parallel computing with
classical computers can be achieved by breaking down the problem to be solved
into discrete parts that can be solved at the same time on different processors
that are linked together. Each processor takes one input, does a computation,
and provides the user with one output. In contrast, a quantum computer is
based on quantum bits or the so-called qubits, which can exist in the states 0
and 1, like a classical bit, but can also take on all values in between. A quan-
tum computer takes several input qubits and performs a computation using a
single quantum processor on all the input qubits simultaneously. Consequently,
a quantum computer has the potential to be millions of times more powerful
than today’s most powerful supercomputers. In fact, it was shown that a 30-qu-
bit quantum computer has a processing power equivalent to a classical com-
puter that could run at 10 teraflops (trillions of floating-point operations per
second) [30]. Our desktop computers today run at speeds measured in gigaflops
(billions of floating-point operations per second). Quantum computers have
the potential to solve complex computational problems conventional comput-
ers are incapable of handling as well as radically changing our cybersecurity
landscape. This is great news for ISR missions where time involved in tracking
and locating adversaries, detecting anomalies, attacking a site or a group of
enemy forces is critical.
 Future Technologies to Enable All-Domain 215

Because of the difficulty in factoring, encryption algorithms to protect in-

formation in naval network traffic are based on factoring large numbers. Factor-
ing a number with 193 digits would take a modest modern desktop computer
with a 2.2-GHz processor 30 dedicated years to factor. A quantum computer
can do the same job in 0.1 seconds.
What then currently impedes the development of a quantum computer?
Why do we not have a quantum computer available for our desks? Part of the
answer depends on the hardware of choice required to implement a qubit and
part of the answer depends on the architecture model. Qubits can be imple-
mented using ions, neutral atoms, color centers in solid-state crystals and super-
conducting circuits. Each implementation has advantages and disadvantages.
Ions can be confined for a very long times [31] and can easily be individually
addressed (read out) but current ion traps are limited to only several ions at a
time [32]. Neutral atoms traps, on the other hand, can hold upwards of a bil-
lion atoms, but individual atom addressing is challenging and the hardware
for atom traps is both costly and complex [32]. Color centers are attractive
because they are embedded in a solid-state crystal that provides for a naturally
occurring vacuum system (in contrast to ions and atoms), but that very lattice
also causes each color center to be slightly different than its neighbors [33]. Fi-
nally, superconducting circuits often need to operate at cryogenic temperatures
[34]. Different approaches to a quantum computer processor also exist, the
most popular being known as the circuit model and the annealing model. The
circuit model follows more traditional computer logic: a qubit is manipulated
with quantum circuits performing the desired operation to achieve the answer.
A quantum algorithm based on annealing casts the problem to be solved as
a minimization problem (e.g., what is the shortest path between two points)
and then finds the global minimum to the problem. A quantum annealing
machine, known as the D-Wave machine, is commercially available. However,
it is based on superconducting qubits and requires operation of the qubits at 17
mK (cryogenic temperatures in millikelvin, a value of 1,000 times smaller than
1 Kelvin). Furthermore, improvement of computing speeds due to quantum
effects has not been proven.
The very principle that gives quantum computers their massively parallel
computing ability also makes the qubit very fragile to manipulate. The super-
position of 0s and 1s needs to be maintained precisely throughout the entire
operation. Ions couple to each other through electrical forces. Atoms are very
sensitive to forces that arise from external potentials, such as electric, magnetic
field, and gravitational fields. Phonons-quantum particles that arise in solid
materials due to vibrations can interact with qubits in solid-state color centers
and destroy the superposition. The challenge to the computing community is
to find qubits that can be rapidly manipulated but are resistant to the effects
of the environment at least during the manipulation time. Furthermore, the
216 Naval ISR Fusion Principles, Operations, and Technologies

architecture needs to scale well such that the equipment required to make a
qubit cannot scale-up unreasonably with the number of qubits. Another chal-
lenge to the community is to demonstrate quantum supremacy, which is the
demonstration of a calculation that can be done either faster or even at all on
a quantum computer versus a classical computer. Once this has been achieved,
the applications to naval ISR can be significant, for example:

• Fast optimization occurs by applying simulated or numerical optimiza-

tion methods to orchestrate sensors to optimize collection in dynamic
environments. Optimizations is required to consider all possible sensor-
target pairs or multisensor-target tuples, over a future time horizon, and
select the set of collections that maximize a global objective function.
While a conventional algorithm searches an unsorted database of N
items in O(N) time, Grover’s search algorithm [35] uses O ( N ) time.
Although this is only a polynomial speed up and not an exponential
one like Shor’s algorithm, it was proven that Grover’s search algorithm
outperforms any classical algorithm and can make a significant differ-
ence in the search time especially when the database consists of a large
number of entries.
• Signal processing of very small signals and predetection fusion of the
raw signals from multiple sensors will benefit from the ability to process
large sets of small-signal data over many sensors to perform correlation
and detection.

10.3.2 Quantum Sensing

As discussed previously, qubits are very sensitive to environmental disturbanc-
es and their fragility can be an impediment to the development of large-scale
quantum computers. However, this disadvantage can be viewed as an advantage
in the field of sensing. The qubit that is sensitive to the environment can be
used to measure the environmental disturbances. This approach is basically at
the heart of the field of quantum sensing.
How might quantum sensors work? To answer this question, we’ll focus
on a subset of quantum sensors known as light pulse atom interferometers,
since atoms can be made to behave like waves and can be manipulated with
laser fields. Atoms can be made to go in two directions at once (a beam splitter),
much in the same way that water waves on a boundary can split and go in two
directions. These waves can be redirected towards each other (an atomic mir-
ror) and then recombined on a beam combiner (a beam splitter in reverse). The
atom has traveled between two points via two different paths and carries infor-
mation about the two paths. By recombining the wave together, the difference
 Future Technologies to Enable All-Domain 217

between the two paths can be measured with extremely high precision set by
the wavelength of the waves. This principle is at the heart of modern-day opti-
cal gyroscopes (which appear in the form of a ring laser gyroscope). The same
principle is at work in an atom interferometer except that the associated wave-
length scales with the mass of the atom. Since an atom weighs more than a
photon [36], an atom interferometer is far more sensitive per atom than an op-
tical gyroscope is per photon. Atom interferometers can be configured to very
sensitively measure forces that arise from gravitational, rotational, or magnetic
potentials. For example, they can be accelerometers, gyroscopes (useful for iner-
tial navigation), or magnetometers (useful for antisubmarine warfare, magnetic
navigation, or explosive detection). The basic operating principle of an atomic
clock (e.g., precision location and high-speed communication) is very similar
to an atom interferometer.
To explore the connection between quantum computers and quantum
sensors, we should discuss the so-called DiVincenzo criteria, which are the char-
acteristics required for a quantum computer, regardless of the architecture or
hardware implementation. Quoting from the original paper, the characteristics
are [37]:

• “A scalable physical system with well characterized qubits,” which means

whatever is used for a qubit’s states 0 and 1 needs to be well defined and
well understood;
• “The ability to initialize the state of the qubits to a simple fiducial state,”
which means the computer needs to be able to set all the qubits into
well-known states (e.g., all 0s);
• “Long relevant decoherence times, much longer than the gate operation
time,” which means the time before the fragile qubits are no longer use-
ful needs to exceed the time for the computation;
• “A ‘universal’ set of quantum gates,” which means that there must be
ways to manipulate the qubits in such a way as to achieve the desired
quantum operation without destroying the qubit;
• “A qubit-specific measurement capability,” which means when an opera-
tion is complete and the qubit holds the answer, it must be able to be
measured.

What does an atom interferometer look like considering the DiVincenzo

criteria?

• Qubits: An atom has very well-defined energy states that can easily be
measured to form the basis of 0s and 1s. These states are the different
218 Naval ISR Fusion Principles, Operations, and Technologies

orbitals that the electrons within an atom can reside in. Alternately, the
qubit can be thought of as the spin of the atom, where the spin is either
up or down (or anywhere in-between).
• Qubit initialization: Using properly constructed laser pulses, the states
of the atom can be manipulated to be initialized so that all the atoms are
in the same state (e.g., all spins down).
• Decoherence times: Different energy states can often decay very rapidly.
However, certain states (either the so-called ground states of the atom or
otherwise very long-lived states) can have long decay times. When atoms
are collected in very low-density samples, the rate of collisions is also
very low, so the decoherence time can be made very large.
• Quantum gates: The very manipulation of the waves as beam-splitters
and mirrors are the analogs of gates. By applying very specially tailored
light pulses, the atomic waves can be split and recombined.
• Measurement: At the end of the interferometer, the state of the atom can
be detected, often by an appropriately tuned laser pulse.

Quantum sensors in general and specifically light pulse atom interferom-

eter sensors have been demonstrated to work in laboratories and are starting to
be demonstrated in more operationally relevant environments [38]. However,
as with any precision sensor, much engineering is required before a sensor can
be developed that fully exploits the fundamental sensitivity of the device. The
intensity, frequency, and polarization of all the lasers used need to be stabilized.
The detection electronics need to be electrically quiet. The number of atoms
used in the source needs to be held constant from measurement to measure-
ment. In a large sense, these requirements are just engineering. But there are
more fundamental issues that also need to be addressed. In principle, the lon-
ger the atoms can evolve untouched in the potential they must measure, the
more precise a measurement can be made but then other factors need to be
considered. It is also for this reason that atom interferometers will be useful
for certain applications and not for others. Atom interferometers can make
excellent gyroscopes for accurately and precisely measuring rotation. But the
precision depends on the length of time the atoms measure the rotation. There-
fore, high dynamic environments that require a rapid read out of the rotation
rate (e.g., a missile during launch) are likely not suitable platforms for an atom
interferometer. However, platforms with low dynamics that need accuracy and
precision for a long time are excellent locations for atom interferometers (e.g.,
submarines that generally turn slowly and need to operate without GPS for
long periods of time).
 Future Technologies to Enable All-Domain 219

10.3.3 Quantum Communication

When we purchase items using credit cards, when warfighters on aircrafts com-
municate with their colleagues on the ground, when we do mission planning
analyses, or when we wish to communicate sensitive information to another
party, we are usually very concerned about the security of our information,
which relies on the assumption that it is computationally hard to break current
cryptographic schemes. For example, the security of the most used public key
encryption, the RSA algorithm, is based on the difficulty of finding the prime
factors of a large number in a reasonable time using the current technology
[39].
However, the development of a large-scale quantum computer capable of
implementing quantum algorithms such as Shor’s algorithm places such crypto-
graphic systems at risk because it will be able to factor numbers in polynomial
time. Consider the difference in time required by a classical and a quantum
computer to factor numbers that are 500 digits long. It is remarkable that fac-
toring a 500-digit number on a classical computer will take more time than the
age of the universe whereas a quantum computer can factor the same number
in two seconds [40]. These numbers suggest that a fully operational quantum
computer will render our currently secure data vulnerable to adversary attacks.
Although the implementation of quantum computers is still at its infancy and
the publicly known quantum computers are either too small (on the order of
few qubits) or not designed to attack our current cryptosystems, the birth of a
fully operating quantum computer may not be too far away given the resources
devoted to this area of research by government agencies, universities, and orga-
nizations throughout the world. It is therefore imperative to adopt new ways of
re-encrypting the existing stored data and ensuring that all forms of communi-
cation are secure; though not unique to naval applications, this is critical for the
highly networked and secure operations of DMO.
Although quantum computers threaten our cybersecurity, quantum theo-
ry upon which it is based can be used to construct cryptographic schemes that
are inherently secure. The essence of secure communication is based on the abil-
ity to establish a secure key that can be used to encrypt and decrypt messages.
Quantum key distribution (QKD) is a technique that uses the laws of quantum
mechanics to generate a secure key among two or more parties over a classical
untrusted channel. The generated key has the property of not depending on
any input value, a task that is impossible to achieve using classical cryptography
[41].
An example of a QKD protocol is the well-known BB84 protocol pio-
neered by Gilles Brassard and Charles Bennett in 1984 [42]. In this scheme,
220 Naval ISR Fusion Principles, Operations, and Technologies

• Alice wants to transmit photons that encode bits of information to Bob

with the goal of creating a shared secret random string of 0s and 1s that
will later be used as a key to encrypt and decrypt messages.
• Alice creates random bits in either a rectilinear or a diagonal basis.
• After Bob makes random measurements in one of these two bases, he
communicates with Alice through a public channel and keeps the bits
where they used the same bases and discard instances when they chose
different bases.
• If an eavesdropper (traditionally named Eve) measures the state of the
photons sent by Alice, she will send photons in the wrong basis 50% of
the time to Bob because the laws of quantum mechanics prohibit her
from making perfect copies of nonorthogonal unknown quantum states.
• When Alice and Bob compare a subset of the random key to test for the
security of their channel, they will notice errors which alert them that
the channel is not secure.

The longest distance over which QKD was demonstrated is 184.6 km

over a fiber-optic link [43]. QKD was also done over a free space link over a
distance of 144 km at a rate of 12.8 bits/second [44]. This free space distance
can allow in principle communication between any two points on earth via a
constellation of orbiting relay satellites [45]. One of the limitations of QKD is
that photons are fragile and therefore suffer from decoherence and noise during
their transmission especially when the distance traveled is long, which lowers
the rate of key generation. This issue will be resolved as research continues in
this rapidly evolving field because QKD is considered an important tool in
cryptography.
There are currently commercial companies that provide QKD devices
such as ID Quantique in Switzerland, MagiQ Technologies in New York, Quin-
tessence Labs in Australia, and SeQureNet in France [46]. In 2018, the group of
Jian-Wei Pan, in collaboration with the Austrian group of Anton Zeilinger, re-
ported on a truly impressive demonstration of intercontinental quantum com-
munication [47]. First, a satellite (named Micius) shared a full QKD protocol
key (including sifting, error correction and privacy amplification) with each of
three ground stations located in Xinglong and Nanshan (both located in China)
and Graz (located in Austria). Next, the satellite was used as a trusted relay to
establish keys among the three ground stations. Using these keys, images were
securely sent from Beijing to Vienna (a picture of Micius) and from Vienna to
Beijing (a picture of the scientist, Schrodinger). Finally, a video conference was
held between the Chinese Academy of Sciences and the Austria Academy of
Sciences, with fiber-based quantum networks connecting the Xinglong station
 Future Technologies to Enable All-Domain 221

to the conference center in Beijing along a 280-km optical fiber link using
six trusted relay stations. The call used the Advanced Encryption Standard
(AES)-128 protocol which refreshed the 128-bit seed every second. During the
75-minute video conference, approximately a total of 2 GB were transmitted,
consuming approximately 70 kB of the quantum key that was generated be-
tween Austria and China. This experiment represents a promising demonstra-
tion of secure video (high bandwidth) transmission using quantum technology,
a feature that can certainly benefit ISR networks.

Endnotes
[1] Cited in the Australian Strategic Policy Institute China University Tracker at https://uni-
tracker.aspi.org.au/universities/naval-research-academy/. The Chinese website is https://
web.archive.org/web/20190724043226/http://navy.81.cn/content/2019-03/15/content_
9450788.htm.
[2] Jones, R.V., The Wizard War: British Scientific Intelligence 1939-1945, NY: Coward, Mc-
Cann & Copenhagen, 1978, p. 73.
[3] Cebrowski, A., Preface to Hughes, W. P., Jr., and R. P. Girrier, Fleet Tactics and Naval
Operations, Third Edition, U.S. Naval Institute Press, 2018.
[4] A similar definition is provided in Handel, M., War, Strategy, and Intelligence, London:
Frank Cass, 1989, p. 113: “a unilateral advantage gained by the introduction of a new
weapon (or by the use of a known weapon in an innovative way) in a war against and ad-
versary who is either unaware of its existence or not ready with effective countermeasures,
the development of which require time.”
[5] “Avoiding Technology Surprise for Tomorrow’s Warfighter: A Symposium Report,” Wash-
ington, D.C.: The National Academies Press, National Research Council, 2009, p. 10.
[6] See, for example, the following reports on technology surprise and reconnaissance-assess-
ment methods:
“Report of the Defense Science Board 2008 Summer Study on Capability Surprise Vol-
ume I: Main Report,” September 2009, Office of the Under Secretary of Defense for
Acquisition, Technology, and Logistics, Washington, D.C., 20301-3140.
“Avoiding Technology Surprise for Tomorrow’s Warfighter: A Symposium Report,” Wash-
ington, D.C.: The National Academies Press, National Research Council, 2009.
“Avoiding Surprise in an Era of Global Technology Advances,” Washington, D.C.: The
National Academies Press, NRC (National Research Council), 2005.
“Persistent Forecasting of Disruptive Technologies,” Washington, D.C.: The National
Academies Press, National Research Council, 2010.
“Technical Assessment: Data-Enabled Technology Watch and Horizon Scanning,” Office
of Technical Intelligence, OASD for R&E, October 2015.
Sylak-Glassman, E. J., S. R. Williams, and N. Gupta, “Current and Potential Use of Tech-
nology Forecasting Tools in the Federal Government,” IDA Document D-5735, Institute
for Defense Analysis.
222 Naval ISR Fusion Principles, Operations, and Technologies

[7] “Strategic Issues Facing Transportation, Volume 3: Expediting Future Technologies for
Enhancing Transportation System Performance,” Washington, D.C.: The National Acad-
emies Press, National Academies of Sciences, Engineering, and Medicine, 2013.
[8] Ibid, p. 2.
[9] Report of the Defense Science Board 2008 Summer Study on Capability Surprise, Volume
I: Main Report, pp. xii, 29.
[10] For example, the U.S. Office of Naval Research identifies its key unclassified technology
areas. See the list of over 100 categories at https://www.onr.navy.mil/our-research/
technology-areas.
[11] For an example, see: “Naval Research and Development: A Framework for Accelerating to
the Navy & Marine Corps After Next,” Department of the Navy, 2018, https://www.onr.
navy.mil/en/our-research/naval-research-framework.
[12] Clark, B., “The Emerging Era in Undersea Warfare,” Center for Strategic and Budgetary
Assessments (CSBA), https://csbaonline.org/uploads/documents/CSBA6117-New-Era-
Undersea-Warfare-Reportweb.pdf.
[13] For an overview of the U.S. Navy strategy and variety of unmanned vehicles, see
Department of the Navy, Unmanned Campaign Framework, March 16, 2021.
[14] “Autonomous Undersea Vehicle Requirements for 2025,” Chief of Naval Operations
Undersea Warfare Directorate, February 2016. See also the earlier RAND study, Button,
R. W., et al., “A Survey of Missions for Unmanned Undersea Vehicles,” RAND MG808,
2009, https://doi.org/10.17226/11927.
[15] Evans, K., “The Nation’s Fixed Undersea Surveillance Assets—A national Resource for the
Future,” The Journal of the Acoustical Society of America Vol. 95, 1994, p. 2852, https://
doi.org/10.1121/1.409559. See also “Origins of SOSUS,” Submarine Fleet Pacific,
Commander, Undersea Surveillance, https://www.csp.navy.mil/cus/About-IUSS/Origins-
of-SOSUS/.
[16] “Department of Defense Authorization for Appropriations for Fiscal Year 2003,”
Underwater Acoustics, pp. 203–204.
[17] Waterston, J., “Ocean of Things: Low-Cost Distributed Sensing with Scalable Analysis for
Maritime Situational Awareness,” DARPA, April 7, 2020. See also Waterston, J., J. Rhea,
S. Peterson, L. Bolick, J. Ayers, and J. Ellen, “Ocean of Things: Affordable Maritime
Sensors with Scalable Analysis,” OCEANS 2019, Marseille, France, 2019, pp. 1–6, doi:
10.1109/OCEANSE.2019.8867398.
The Ocean of Things program is a research and development effort that aims to provide
the expanded information capabilities associated with the Internet of Things to the
ocean in a cost-effective way, taking advantage of developments in cloud-based analytic
techniques to assess maritime conditions and contribute to oceanographic models in near
real-time. The objective of the program is to provide environmental sensing and maritime
surveillance missions by developing new low-cost methods for detection and tracking of
objects and characterization of oceanographic phenomena at a fine scale over broad areas.
The inputs for these data analytics over a large ocean area would be provided by deploying
thousands of small, low-cost smart floats to form a distributed sensor network. Source:
Ocean of Things Broad Agency Announcement, HR001118S0013, DARPA Strategic
Technology Office, December 19, 2017.
 Future Technologies to Enable All-Domain 223

[18] Martin, D. L., Autonomous Platforms in Persistent Littoral Undersea Surveillance:

Scientific and Systems Engineering Challenges, October 6, 2005.
[19] Source: ONR APEX Undersea Challenge 2021, London Tech Bridge, https://www.onr.
navy.mil/londontechbridge/Statements.
[20] Figure adapted from Fletcher, B., “New Roles for UUVS in Intelligence, Surveillance, and
Reconnaissance,” Space and Naval Warfare Systems Center, November 2000. Figure from
Brininstool, M., “Undersea Fiber Optic Technology,” briefing, Space and Naval Warfare
Systems Center, San Diego, CA, 92152.
[21] Ball, M., New Open Standard for Secure Underwater Communications, Sonardyne,
June 27, 2021, https://www.unmannedsystemstechnology.com/2021/06/new-open-
standard-for-secure-underwater-communications/.
[22] Gussen, C. M. G., et al., “A Survey of Underwater Wireless Communication Technologies,”
Journal of Communication and Information Systems, IEEE, Vol. 31, No. 1, 2016,
pp. 242–254.
[23] Yan, Z., Xu, D., and Chen, T., Formation Control of Unmanned Underwater Vehicles
Using Local Sensing Means in Absence of Follower Position Information, International
Journal of Advanced Robotic Systems, January 27, 2021, https://journals.sagepub.com/doi/
full/10.1177/1729881420986745.
[24] “Distributed Remote Sensing for Naval Undersea Warfare (Abbreviated Version),”
Washington, D.C.: The National Academies Press, National Research Council 2007,
pp. 3–4.
[25] L. Martin, D. L., ibid.
[26] Moore, G. E., “Cramming More Components onto Integrated Circuits,” Electronics,
Vol. 38, 1965 p. 114.
[27] Feynman, R. P., “Simulating Physics with Computers,” International Journal of Theoretical
Physics, Vol. 21, No. 6/7, 1982, pp. 467–488.
[28] Shor, P. W., “Polynomial-Time Algorithms for Prime Factorization and Discrete
Logarithms on a Quantum Computer,” SIAM Journal on Computing, Vol. 26, No. 5,
1997, pp. 1484–1509.
[29] “Quantum Computing: Progress and Prospects,” Washington, D.C.: The National
Academies Press, National Academies of Sciences, Engineering, and Medicine 2019,
pp. 184–185, https://doi.org/10.17226/25196.
[30] Sarma, K. J., “Understanding Quantum Computing,” International Journal of Scientific
Engineering and Applied Science, Vol. 1, No. 6, 2015, pp. 370–388.
[31] Bruzewicz, C. D., J. Chiaverini, R. McConnell, and J. M. Sage, “Trapped-Ion Quantum
Computing: Progress and Challenges,” Applied Physics Review, Vol. 6, 2019, p. 021314.
[32] Bradley, C. C., and R. Hulet, Laser Cooling and Trapping of Neutral Atoms, Experimental
Methods in the Physical Sciences, Vol. 29, 1996, pp. 129–144.
[33] Doherty, M. W., N. B. Manson, P. Delaney, F. Jelezko, J. Wrachtrup, and L. C. Hollenberg,
“The Nitrogen-Vacancy Colour Centre in Diamond,” Physics Reports, Vol. 528, No. 1,
2013, pp. 1-45.
224 Naval ISR Fusion Principles, Operations, and Technologies

[34] Xiang, Z.-L., S. Ashhab, J. Q. You, and F. Nori, “Hybrid Quantum Circuits:
Superconducting Circuits Interacting with Other Quantum Systems,” Reviews of Modern
Physics, Vol. 85, 2013, p. 623.
[35] Grover, L. K., “Quantum Mechanics Helps in Searching for a Needle in a Haystack,”
Phys. Rev. Lett., Vol. 79, 1997, p. 325. See also Boyer, M., G. Brassard, P. Hoyer, and A.
Tapp, “Tight Bounds on Quantum Searching,” Fortsch. der Phys. Prog. of Phys., Vol. 46,
No. 493, 1998.
[36] Of course, a photon is massless, but it has an effective mass given by E = h c / l = mc2, where
l is the wavelength, c is the speed of light, h is Planck’s constant, and m is the effective
photon mass.
[37] DiVincenzo, D. P., “The Physical Implementation of Quantum Computation,” Fortschritte
der Physik, Vol. 48, No. 9-11, 2000 pp. 771–783.
[38] Narducci, F. A., A. T. Black, and J. H. Burke, “Advances Toward Fieldable Atom
Interferometers,” Advances in Physics: X, Vol. 7, No. 1, 2022, DOI: 10.1080/
23746149.2021.1946426.
[39] Ambedkar, B., and S. Bedi, “A New Factorization Method to Factorize RSA Public Key
Encryption,” IJCSI International Journal of Computer Science Issues, Vol. 8, No. 6, 2011
p. 1.
[40] Smith, D., “Notes from the Back Row: ‘Quantum Entanglement and Quantum
Computing’,” 2013, https://www.caltech.edu/about/news/notes-back-row-quantum-
entanglement-and-quantum-computing-39378.
[41] Stebila, D., M. Mosca, and N. Lutkenhaus, “Quantum Communication and Quantum
Networking: First International Conference,” Naples, Italy, 2009.
[42] Bennett, C. H., and G. Brassard, “Quantum Cryptography: Public Key Distribution and
Coin Tossing,” in Proceedings of IEEE International Conference on Computers, Systems, and
Signal Processing, 1984.
[43] Hiskett, P. A., D. Rosenberg, C. G. Peterson, R. J. Hughes, S. Nam, A. E. Lita, A. J.
Miller, and J. E. Nordholt, “Long-Distance Quantum Key Distribution in Optical Fibre,”
New Journal of Physics, Vol. 8, 2006, p. 193.
[44] Schmitt-Manderbach, T., et al., “Experimental Demonstration of Free-Space Decoy-State
Quantum Key Distribution over 144 km,” Phys. Rev. Lett., Vol. 98, 2007 p. 010504.
[45] Stebila, D., M. Mosca, and N. Lutkenhaus, ibid.
[46] https://www.quantiki.org/wiki/private-sector-quantum-information-science.
[47] Liao, S.-K., et al., “Satellite-Relayed Intercontinental Quantum Network,” Phys. Rev.
Lett., Vol. 120, 2018, p. 030501.
About the Authors
Jim Scrofani is the director of the Center for MultiINT Studies (CMIS) at the
Naval Postgraduate School (NPS) where he has worked as an associate professor
since 2011. A United States Navy Captain, he was previously responsible for
program management of an extensive research and development (R&D) port-
folio (2009–2011) at the National Reconnaissance Office, Advanced Systems
and Technologies directorate. He was also the senior naval space systems engi-
neering duty officer Space and Naval Warfare Systems Command (SPAWAR)
space field activity and served as senior community mentor and certifying of-
ficer for all space-qualified and designated naval engineering duty officers. His
prior Navy experience includes assignments with the Multi-National Security
Transition Command–Iraq as commander, operations chief; National Recon-
naissance Office, as advanced systems and technologies directorate, command-
er, and program manager responsible for the execution of classified space and
ground-based advanced technology demonstrations; and the Navy, Strategic
Systems Program, responsible for all in-factory rocket motor development, pro-
duction, and operational support of the Navy’s Trident II missile systems. In
addition, he has served as the assistant for arms control to the director, Strategic
Systems Programs, responsible for all aspects of implementation and compli-
ance with the Strategic Arms Reduction Treaty (START). From 1989 to 1994
he served as a submarine officer assigned to Commander, Submarine Group 10
(COMSUBGRU TEN), serving aboard the USS Spadefish (SSN-668). He has
earned four degrees: a PhD in electrical engineering from the Naval Postgradu-
ate School, 2005, an MS with distinction in electrical engineering from the
Naval Postgraduate School, 1997, an MBA from Brenau University in 1994,
and a BS in chemical engineering from the University of Florida, 1987.

225
226 Naval ISR Fusion Principles, Operations, and Technologies

Will Williamson is currently an associate research professor in the Elec-

trical and Computer Engineering Department of the NPS, where he supports
the NPS Center for MultiINT Studies in the Washington, D.C. region. His
current research interests are in applying machine learning and artificial intel-
ligence to various defense and intelligence applications, including cyber intru-
sion detection, characterization of social media influence operations, distrib-
uted cognition for small satellite constellations, and multisensor/multiplatform
ISR systems. Dr. Williamson began his career as a helicopter pilot in the Marine
Corps and Army National Guard, with corollary duties as logistics officer and
intelligence officer. He earned his PhD in experimental condensed matter phys-
ics from the University of Toledo while working full-time in support of Defense
Advanced Projects Agency (DARPA) programs at the Special Purpose Processor
Development Group for the Mayo Foundation. Dr. Williamson spent the next
25 years supporting the DoD and the intelligence community. He has worked
extensively with DARPA in a variety of roles over 15 years. He has worked for
the Center for Naval Analysis, Institute for Defense Analysis, DARPA, and MI-
TRE Corporation on research projects for each of the armed services and sev-
eral intelligence agencies. He has worked on multisensor fusion for detection of
naval mines, landmines, and other tactical and strategic targets, and has worked
on sensor technologies including lidar and vertically interconnected sensor ar-
rays. He has also worked extensively in cognitive systems to support dynamic
sensor tasking. Dr. Williamson is an adjunct professor of strategic intelligence
at Patrick Henry College.
Jihane Mimih is the CMIS director of programs and a research assis-
tant professor at the NPS. Since 2017, she has directed research programs in
CMIS and conducted research in the areas of maritime domain awareness,
anomaly detection, data fusion, quantum sensing, and computing. As director,
she oversees program development, center business operations, and execution
of research and academic programs in intelligence integration, space systems,
data fusion, and technology reconnaissance. She has organized and led focused
colloquia and meetings comprised of faculty members, industry, intelligence
agencies, and project managers to identify viable sources of data that can be
used to enhance key research efforts. Prior to 2017, Dr. Mimih was a physi-
cist at the Naval Air Warfare Center, Patuxent River, Maryland, antisubmarine
warfare division, where she conducted theoretical physics research to solve out-
standing antisubmarine warfare (ASW) problems using nonacoustic tools. She
performed basic research in quantum optics and quantum information with
the goal of improving the sensitivity of quantum sensors and developing quan-
tum key distribution protocols. She founded and led the Quantum Computing
Working Group at NAVAIR to develop a workforce fluent in the capabilities of
quantum computing.
 About the Authors 227

Ed Waltz is professor of practice (intelligence) in the CMIS at the NPS

where he conducts research in intelligence processing from multiple intelligence
sources. Prior to joining NPS in 2018, he was division chief, advanced concepts
at the National Reconnaissance Office, while a distinguished member of the
technical staff at Virginia Tech, serving an IPA position within the U.S. intel-
ligence community, where he is the division chief, leading intelligence process
research focused on automating intelligence collection and data integration.
Prior to 2013, he held numerous research leadership roles at BAE Systems Ad-
vanced Information Technologies (chief scientist), General Dynamics, and Ve-
ridian (technical director, senior scientist)—developing and deploying signal
processing, data fusion, and intelligence analysis capabilities. He led numer-
ous hard target MultiINT studies and tool developments over two decades for
different agencies of the Intelligence Community (IC). His prior experience
was as a digital signal processor designer (Apollo, NASA multispectral scanner,
LANDSAT), systems engineer (Navy radiometer, SEASAT, space shuttle), and
systems engineering manager (AN/UPX-30(V), Tri-Service Mark XV). Since
2015, he has been an adjunct professor of strategic intelligence at Patrick Henry
College. He has given over 45 international lecture series (2- and 3-day courses)
on intelligence topics from 1990 to 2010.
Index
ACINT, 28 Center for International Maritime Security
Acoustic communications (ACOMM), 136, (CIMSEC), 25
207, 210 Chemical, biological, nuclear, radiological,
Activity, 115–118, 124, 165 and explosive (CBNRE), 32
Activity based intelligence (ABI), 118 Combat Information Center (CIC), 69
Advanced persistent threats (APT), 103, 104 COMINT, 27
Air and missile defense (AMD), 10, 62 Commander’s intelligence requirements
Algorithmic warfare, 157 (CIRs), 125
All-domain Common operating picture (COP), 8, 69,
fusion, 87–90, 113 89,113,133–135
ISR, 5, 7,8, 172 Computer network exploitation (CNE), 104
All-source intelligence analysis, 108 COMSATCOM, 137
Antisatellite (ASAT), 37 Concept of Operations (CONOP) ,
Antiship missile (ASM), 56, 92 84,138,145
Automatic identification system (AIS), 28, Consensus fusion system, 166
38–40, 88, 89, 93, 96, 115–117, Constellation
182, 183, 185 Capella SAR, 181–182
Automation, 12, 66, 69, 90, 107, 123, 155, Planet Labs EO, 180–181
157–161, 165–172 Hawkeye RF signals, 182–183
Autonomy, 155, 207, 211 performance, 187–193
Artificial intelligence and machine learning transport layer, 184–185
(AI/ML), 3, 29, 58, 155–168 Cooperative engagement capability (CEC),
Antiaccess/area denial (A2AD) 6, 56–57 23, 61, 62, 98–100
Antisubmarine warfare (ASW), 10, 3, 62, Course of action (COA), 14, 121–122, 162,
71, 131 171
Cyber ISR, 27, 42, 104
Battle force manager (BFM), 64 Cyber operational preparation of the
environment (C-OPE), 27, 43
C2 agility, 169, 172 Cyberspace operations (CO), 27, 40, 72
C2 automation, 167–172
C4ISR, 2, 82 Data fusion, 26, 108, 109, 134, 144, 146
all-domain, 87, 113

229
230 Naval ISR Fusion Principles, Operations, and Technologies

Data fusion (continued) Fifth Generation (5G) wireless

centralized, 113, 134, 146 networks, 137–139
distributed (DDF), 144–152 FORCEnet, 79, 133
hierarchical, 146 Forward Pass, 99–100
model-based, 115 Functional nodes, 139–143
Data fusion levels
level 1, entity assessment, 16 Geospatial intelligence (GEOINT), 26, 28
level 2, situation assessment, 16, 76–77, Geosynchronous Earth orbit (GEO), 36
81, 112, 114, 115 Global information grid (GIG), 72
level 3, impact assessment, 16, 112, 120 Gray zone, 6
level 4, performance assessment, 16 Great power competition, 5, 6
Data Ground sample distance (GSD), 38
hard, 162
pedigree, 113 Heavy weight UUV vehicles (HWV), 32
provenance, 113 Highly-elliptical Earth orbit (HEO), 36
soft, 162 Human intelligence (HUMINT), 28, 141,
Data mining, 17–18 162
Defensive counter cyber (DCC), 43 Human machine interface (HMI), 16
Denial of service (DoS), 105 Hypersonic, 6, 37, 64, 95, 178, 202, 205
Directed energy, 36, 37, 201
Distributed
Indications and warnings (I&W), 44, 107
detection, 132
Information dominance, 5, 187
identification, 132
Infrared (IR), 38
lethality, 5, 24, 57, 103
Integrated fire control (IFC), 98
sensing, 91, 147
Intelligence Surveillance and Reconnaissance
sensor network (DSN), 131–132
(ISR), 1–14, 28
tracking, 132, 146–152
national, networked, 9
Distributed Common Ground System-Navy
organic, 9, 46
(DCGS-N), 78
Intelligence, surveillance, reconnaissance,
Distributed Maritime Operations (DMO),
and targeting (ISRT), 40, 46–51,
5, 11–12, 23–24, 55–68, 90–103,
87–88, 141
123, 133, 169, 206
International Maritime Organization
architecture, 60–62
(IMO), 25
introduction, 56–58
scenario, 64–66
tactical considerations, 58–60 Joint All-Domain Operations (JADO), 28,
Dynamic force employment (DFE), 5, 24 87, 107
Joint Directors of DoD Laboratories (JDL),
14, 109–114, 119
Electronic intelligence information
Joint Force Maritime Component
(ELINT), 27
Command groups (JFMCC), 78
Electronic warfare (EW), 37, 71
Joint Intelligence Preparation of Operating
Electro-optical (EO), 4, 38, 38–39, 165, 192
Environment (JIPOE), 47, 121,
Emission control (EMCON), 38, 120
122
Engage on Remote, 99–100
Enhanced mobile broadband (EMBB), 137
Entity, 119, 145, 162–165 Kalman filter, 148–150
Expeditionary advanced base operations
(EABO), 44, 49 Landing helicopter assault (LHA), 56
Landing helicopter dock (LHD), 56
Landing platform dock (LPD), 56
 Index 231

Landing ship dock (LSD), 56 Object-based production (OBP), 114, 116

Latency, 12, 89, 92, 126, 138, 139 Observe-orient-decide-act (OODA), 12,
Law of the Sea, 2 13–15, 157, 158
Light weight UUV vehicles (LWV), 32 Offensive counter cyber (OCC), 43
Link-16, 62, 136 OpELINT, 27
Link-22, 136 Open-source intelligence (OSINT), 26, 28,
Links (data), 104, 133–136, 184, 190, 162
209–211 Operational intelligence (OPINTEL), 77
Long-range antiship missile (LRASM), 61 Orbital threats, 37
Low-Earth orbit (LEO), 35, 36, 39, 62, 92, Orchestration, sensor, or collection, 123, 124
136, 139–142 Order of battle, 43, 46
Over-the-horizon (OTH), 57
Machine learning (ML), 16, 58, 117, 158
Marine expeditionary units (MEU), 56 Persistent engagement, 104
Maritime domain awareness (MDA), 2, 5, Persistent surveillance, 8, 205
25, 30, 31, 118 Position in time (POSIT), 71, 89, 93, 94,
Maritime domain, 2, 7, 10, 12, 108, 119 116, 157
Maritime intercept operations (MIO), 10 Process noise, 148
Maritime operational commander (MOC), Publicly available information (PAI), 28,
64, 78 120, 157
Maritime Operations Center (MOC), 14
Maritime security operations (MSO), 10, Quantum communications, 219–221
14, 44 Quantum computing, 214–216
Maritime superiority, 4, 8, 23, 59, 121 Quantum sensing, 216–218
Maritime zones, 3 Quantum technologies, 213–221
Massive machine-type communications
(MMC), 138 Radio frequency (RF), 38, 81, 88, 136, 182,
Measurement and signature intelligence 193, 210
(MASINT), 28, 205 Remote Fire, 100
Medium-Earth orbit (MEO), 36, 136 Report association, 92–93
Mesh network, 88, 134–135, 142, 146, 184 Ring network, 134
Meteorological and oceanographic systems
(METOC), 70, 71, 78, 178
Salvo, 102
MILSATCOM, 137
Seabed advanced deployable system (ADS),
Mine warfare (MIW), 10
33, 205
MultiINT, 12, 79, 107–127, 111, 147,
Seabed fixed surveillance system (FSS), 33
159–161, 191–193
SEALINK, 14
collection, 159–161
Sensemaking, 156, 161–172
Service-oriented architecture, 71, 73
National technical means (NTM), 9 Signal intelligence (SIGINT), 27, 105, 171
NATO, 25, 167, 168 Single-shot kill probability (SSPK), 102
Naval operational architecture (NOA), 133 Situation, 115
Network centric warfare (NCW), 11, Situation awareness, 108, 115, 121
132–133 Situation detection, 115
Network-centric operations, 34, 132–133 Smallsat constellations, 179–183
Nodes, 89, 92, 99, 103–104, 112–113, Social network analysis (SNA), 119
132–134, 139–161 SONAR, 32–33, 50, 141, 205–208
Space Mission Analysis and Design
Object, 118, 124, 132, 157, 165 (SMAD), 189
232 Naval ISR Fusion Principles, Operations, and Technologies

Specific emitter identification (SEI), 88 UML sequence of services, 79–82

Star network, 134 Undersea and seabed surveillance, 205–213
Surface action group (SAG), 44–48, 57, 142, Unified Modeling Language (UML), 79
143, 171 Unmanned
Swarm, 7, 208 air vehicles (UAV), 48, 96, 99, 171,
Synthetic aperture radar (SAR), 38, 39, 50, 211
181, 192, 193 aircraft system (UAS), 57, 62, 104
Seabed to space, 3, 23, 42, 72, 82, 91, 112 surface vessels (USV), 34, 61, 77, 99,
Seabed, 3, 31, 42, 72, 91, 205 171, 208
Surface warfare (SUW), 10, 34, 44, 47, 62 underwater vehicles (UUV), 32, 48–50,
Sea control, 5, 23, 24 211, 212

Tactical data line (TDL), 23, 69, 131, 141 Vertical launch system (VLS), 61
Tactical intelligence (TACINTEL), 32, 77 Vessel of interest (VOI), 14
Target-based intelligence, 118 Visit, board, search, and seizure (VBSS), 10,
TechELINT, 27 14, 15
Technology scanning, 201–204
Technology watch, 199–201, 205 Web Ontology Language (OWL), 158
TENCAP, 9 Wireless networks, 137–139
Track, tracking, 10–12, 38–40, 88–89,
93–104, 110–126, 132, 164–166,
171, 184–186, 205–208