Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Robotics and Computer-Integrated Manufacturing

journal homepage: www.elsevier.com/locate/rcim

Ubiquitous manufacturing: Current practices, challenges, and

opportunities
Toly Chen a, Horng-Ren Tsai b,n
a
Department of Industrial Engineering and Systems Management, Feng Chia University, No. 100, Wenhua Rd, Taichung City 407, Taiwan
b
Department of Information Technology, Ling Tung University, No. 1, Lingtung Rd, Taichung City 408, Taiwan

art ic l e i nf o a b s t r a c t

Article history: Ubiquitous manufacturing (UM) features a “design anywhere, make anywhere, sell anywhere, and at any
Received 29 October 2015 time” paradigm that grants factories an unlimited production capacity and permanent manufacturing
Received in revised form service availability. However, the research and applications of UM have been limited thus far to in-factory
28 December 2015
operations or logistics. For this reason, this study reviews the current practices of UM, discusses the
Accepted 6 January 2016
challenges faced by researchers and practitioners, and determines potential opportunities for UM in the
near future. Finally, we conclude that the success of UM depends on the quality of the manufacturing
Keywords: services deployed, and that UM is a realizable target for Industry 4.0.
Ubiquitous computing & 2016 Elsevier Ltd. All rights reserved.
Ubiquitous manufacturing
Cyberphysical system
Sensor
Industry 4.0

1. Introduction Foust [21] listed four prerequisite conditions for UM:

Ubiquitous computing (UC), or pervasive computing, is a concept in (1) Transporting finished goods is more expensive than trans-
software engineering and computer science in which computing is porting raw materials.
performed at any location [48]. Ubiquitous manufacturing (UM) is an (2) Finished goods are larger, more fragile, or more perishable
application of UC in the manufacturing sector that features a “design than raw materials.
anywhere, make anywhere, sell anywhere, and at any time” paradigm (3) Raw materials are available ubiquitously.
[19,36]. However, manufacturing a product ubiquitously is not easy. (4) Manufacturers and their customers communicate closely.
For this reason, UM typically implied that products could be supplied
ubiquitously [21]. Nevertheless, with the advances in information, These are also called the location controls of UM.
communication, sensing, and networking technology, manufacturing The dynamism and globalization of demand is considered the
services can now be provided ubiquitously [50]. driving force behind the emergence of CM and UM [36]. Micro-
A concept similar to UM is cloud manufacturing (CM), which electromechanical and radio frequency systems were considered
enables ubiquitous, convenient, on-demand network accesses to a key technologies to UC and UM [35]. Under Web 3.0, a program
shared pool of configurable manufacturing resources [50]. In many could run on any device. Thus, Web 3.0 and 4.0 also contributed to
studies, UM and CM have not been differentiated [36]. However, in the realization of UM [23]. According to Strassner and Schoch [41],
contrast to CM, UM emphasizes the mobility and dispersion of automatic identification, localization, and sensor technologies
manufacturing resources and users. In addition, CM typically de- were essential to developing a UC and UM application. Real-time
ploys cloud services through the Internet, whereas UM does not visibility and interoperability were considered core characteristics
necessarily rely on a network. For example, Kinect, a sensing of UM [53].
system widely applied in UM, does not require network access to Ferreira et al. [23] asserted that CM and UM grants factories
operate. Nevertheless, UM and CM benefit from each other. For unlimited production capacity and permanent manufacturing
example, the status of a machine monitored by a ubiquitous sensor service availability. In addition, UM was considered capable of
can be analyzed using a cloud service. sustaining the needed agility and quickness to react to market
changes [37]. UM also frees operators from error-prone and time-
n
Corresponding author. consuming data collection tasks [53]. However, research into and
E-mail address: [email protected] (H.-R. Tsai). applications of UM have been limited [26]. Even the most common

http://dx.doi.org/10.1016/j.rcim.2016.01.001
0736-5845/& 2016 Elsevier Ltd. All rights reserved.

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
2 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎

UM application, radio frequency identification (RFID), was not work-in-process (WIP) management system for a small flexible
common in many regions [52-56]. For this reason, this study re- manufacturing system (FMS) by using smart objects such as RFIDs
views the current practice of UM, discusses the challenges faced by and auto IDs and Web services. The FMS was composed of three
UM researchers and practitioners, and determines possible op- workstations, one trolley, and one shelf. RFID tags were used to
portunities for UM in the foreseeable future. identify operators, components, pallets, and locations on the shelf.
The remainder of this paper is organized as follows. Section 2 RFID readers were integrated with a smart gateway and wrapped
defines UM and UM systems, and introduces some existing UM with Web services to be easily invoked. Thus, the material flows in
systems. Section 3 classifies the technologies for establishing a UM the FMS could be automatically traced; the WIP level could be
system into several categories, along with their applications. The monitored, and, based on the monitoring results, proper shop floor
problems and challenges facing existing UM systems are raised in control actions could be taken.
Section 4, along with their possible solutions. Section 5 discusses UC technologies have been applied to other places in a factory.
potential developments in this field and concludes this paper. For example, Bose and Pal [6] installed auto-ID readers at point-of-
sale, storage, and receiving locations to automate data collection.
During their investigation, several concerns were raised regarding
2. UM systems whether an auto-ID application could be successful, including the
acceptable initial investment, item-level or pallet-level tagging,
2.1. Definitions of UM data storage, analysis, privacy, big-band or phased adoption, and
integration with existing management information systems
Various means of defining a UM system exist [36]. The first is to (MISs).
apply UC technologies to a manufacturing system. Huang et al. In logistics, Prasanna and Hemalatha [34] used RFID tags to
[25] and Zhang et al. [52-56] have defined a UM system as a identify goods to avoid misplacement, and weight sensors to avoid
wireless sensor network that uses RFID tags and receivers, auto- the overload of vehicles. They also used a global positioning sys-
matic identification (auto ID) sensors, and wireless information or tem (GPS) and the GSM to track the locations of vehicles. The
communication networks such as WiFi or the global system for weight sensors were connected to a PIC 16F877A microcontroller.
mobile communications (GSM) to automatically collect, synchro- The GPS receivers were also connected to the microcontroller with
nize, and/or process manufacturing data. By this definition, UM is SN74HC00 gate integrated circuits. Thus, heterogeneous informa-
the same as wireless manufacturing [25] or e-manufacturing [8]. tion sources were aggregated to facilitate decision making.
UM defined in this fashion is typically confined to the operations
of a factory or logistics. 2.2.2. UM systems that deploy manufacturing facilities ubiquitously
Another definition of UM is to deploy manufacturing resources, An experimental LEGO ubiquitous assembly system was es-
services, and/or facilities that use the same raw materials and tablished by several universities and institutes; in it, a customer
produce comparable finished goods as ubiquitously as possible browsed an online catalog and placed an order that could be ful-
[21], that is, the so-called ubiquitous industries. The focus is on filled by any of the participating universities or institutes [35].
internationalization and distributivity [35]. This definition of UM Despite the limited attempts, manufacturing facilities have only
has often been mentioned by geographers. Foust [21] provided an been deployed ubiquitously in the newspaper publishing, bakery,
ad hoc definition of UM as a type of manufacturing that is market- and dairy industries.
oriented and has a frequency of occurrence greater than a specific The ease of installing and using a three-dimensional (3D)
threshold that can be empirically verified as ubiquitous. Specifi- printer and the convenience of exchanging a 3D model online are
cally, UM indicates that more than n production locations exist cultivating additive-manufacturing-based UM systems. For ex-
ample, websites like My Mini Factory (www.myminifactory.com)
within a country for manufacturing a product. Alexandersson [1]
and shapeways (www.shapeways.com) gather 3D models from
set the threshold for classification as ubiquitous as facilities that
volunteers worldwide, such that a 3D model can be downloaded
can be found in all regions with a population of at least 10,000; as
from the website and printed out at any place and time. By con-
such, only two industries in the United States, printing and pub-
trast, websites such as 3D Hubs (www.3dhubs.com) work with 3D
lishing and food processing, satisfied his requirement. A critical
printing facilities worldwide; customers can determine the 3D
concern for such a definition is the balance between the locational
printing facility closest to their location and print the 3D model
costs of production and the distribution costs of finished goods. To
they select from (or upload to) the website.
achieve UM, a hypersized manufacturing network is typically re-
quired [36]. An alternative is to form a UM network system
2.2.3. UM systems based on CPSs and web (and cloud) services
through cross-factory or cross-enterprise collaboration.
MacGregor Cranes, a crane manufacturer, designed and in-
Putnik et al. [35] defined a UM system as a collection of in-
stalled a remote diagnostic system onto each of its cranes to
telligent devices, logically and/or spatially distributed, that change
provide predictive maintenance services to customers [26]. The
dynamically and reconfigure automatically for new tasks through
remote diagnostic system included sensors to monitor the tem-
the provision of manufacturing services, and are supported by
perature, pressure, speed, and usage time of a crane, which were
semantic tools for unambiguous communication. This combines
transmitted to MacGregor Cranes (or another service provider) via
the previous two definitions, with a focus on a cyberphysical
GSM or a satellite modem. Urgent problems or errors could be
system (CPS)1 and Web (or cloud) services.
discovered by occasional readings of the different parameters. To
prevent a potential breakdown, an analysis of long-term data was
2.2. Existing UM systems required. Two ways of applying the remote diagnostic system,
creating different values for a factory, were also compared by
2.2.1. UM systems based on UC applications Jonsson et al. [26]. Conversely, data security and privacy, de-
This section reviews some representative cases of the applica- personalization of the service experience, and error-prone sensors
tion of UC technologies. Zhang et al. [53] established a real-time with a limited capability were some of the shortfalls of the remote
diagnostic system.
1
A CPS is a system of collaborating computational elements controlling phy- To apply cloud services to assist in the predictive maintenance
sical entities [28]. of a machine, information sharing and interactions between the

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 3

sensing system and a cloud service provider (CSP), and between being held to accommodate for the progress of other items.
the CSP and cloud services, are critical. To address this, Wang et al. Through this application, critical information on the progress of all
[46] proposed an agent-based approach in which mobile agents items could be collected using RFIDs. Zhang et al. [59] equipped a
were responsible for distributing the analysis algorithm to cloud trolley with an RFID reader, an antenna, and an industrial control
services. The system was established on a server running the Linux computer to enable active and direct communication regarding
operating system, mobile agent middleware, and open source the pools of transportation tasks.
numerical libraries. Kinect is a sensing system composed of a red–green–blue color
camera, 3D depth sensor, and directional microphone that work
collaboratively to track the gestures and motions of a user from a
3. UM technologies distance [33]. Kinect has been applied to UM. For example, the
gesture or motion of an operator has been detected by Kinect and
3.1. Ubiquitous sensors imitated by an industrial robot to train the robot [18]. The detected
gesture and motion can also be mapped to a computer numerical
Ubiquitous sensors are sensors that support UC (and UM). An code (CNC) to control a CNC machine [14]. Some applications of
ordinary sensor can be turned into a ubiquitous sensor by con- Kinect to manufacturing are summarized in Table 2. The ad-
necting it to a networking module. Cha et al. [8] identified three vantages of applying Kinect to manufacturing include equipping
means of doing this: directly interface a sensor to a fully auto- machines with human intelligence, distance control, training, and
mated device controller through transmission control protocol/ machine operation, thereby enabling machines to replace opera-
internet protocol or serial communications, add necessary per- tors in dangerous or difficult-to-reach manufacturing environ-
ipherals to a sensor and interface it with a device controller ments and elsewhere. By contrast, the disadvantages associated
through transistor–transistor logic or a 4–20 mA current loop, and with Kinect applications include the requirement for program-
enter data manually using a bar-code reader or touch screen. ming to analyze the collected data, a delay (or time lag) in du-
Ubiquitous sensors and sensing technologies include RFID, auto ID, plicating an operator's motions, and limited and overly focused
Kinect, virtual reality (VR), CPS, GPS, GSM, and WiFi. Smart objects applications.
are homogeneous or heterogeneous objects equipped with ubi- VR is the use of information technology (IT) to create the illu-
quitous sensors [52-56]. A smart object can communicate the in- sion that a user is in a particular environment [45]. VR can be
formation it owns [5]. applied to simulate a robotic arm to determine possible collisions
RFIDs and auto IDs have been extensively applied in manu- with the environment, or to simulate a factory that requires a fa-
facturing-related fields. For example, Cheng et al. [15] used RFIDs cility relayout. VR, mixed reality, and augmented reality have been
to collect real-time information on machine availability, WIP level, applied to UM. For example, [30] captured the gestures and mo-
and job status for a shop floor supervisor to make dynamic deci- tions of an operator by using Kinect, and then established a 3D VR
sions. This information came from three types of heterogeneous system to show the captured gestures and motions to ensure that
data sources and services: enterprise information systems, shop they could be fully imitated by a robotic arm. An operator is not
floor databases, and Web services. Some representative cases of confined to a fixed position to be detectable by the Kinect or VR
RFID and auto-ID applications are provided in Table 1. The ad- system, thereby establishing a UM environment. However, estab-
vantages of applying RFIDs and auto IDs to manufacturing include lishing a VR system requires extensive IT knowledge. In addition,
automatic and real-time object tracking, more accurate and timely even if a VR system resembles a factory, it may not function ex-
data, the streamlining and automating of operations, cost reduc- actly like a factory.
tions, increased asset visibility, improved responsiveness, and ex- One of the mainstream applications of CPSs is predictive
tended business opportunities [5]. However, one disadvantage is maintenance, in which various sensors are used to monitor the
the difficulty to establish a flexible and scalable system archi- status of a machine (i.e., live-condition monitoring). The results
tecture to integrate such applications with existing business pro- are then analyzed manually or by a Web (or cloud) service to
cesses [15,5]. A novel attempt to integrate with a factory sche- predict and prevent potential breakdowns [26,8]. Automatic de-
duling system was undertaken by Zhang et al. [54], through which tection also prevents an operator from missing, misinterpreting, or
information regarding the use and current status of each machine incorrectly communicating a machine's failure [26]. However,
was collected by an RFID and submitted into the scheduling sys- connecting a machine to the Internet is risky from the perspective
tem in real time to facilitate accurate rescheduling of jobs in the of data or equipment security. For many factories, a major concern
factory. Zhang et al. [58] applied the RFID technology to an un- when transferring operations to cloud or Web services is the
paced assembly line on which a workpiece was transferred possible leakage of confidential operational data [10]. For example,
whenever the required operation was completed, rather than Karnouskos [29] described how the Stuxnet worm attacked an

Table 1
Representative cases of applying RFIDs and auto IDs to manufacturing.

Reference Field Applications

Bornhövd et al. [5] Logistics 1. Established an auto-ID infrastructure.

2. Tracked the delivery of goods.
Bose and Pal [6] Supply chain management 1. Tracked the flow of materials in a supply chain.
Wong et al. [49] Product authentication –
Zhang et al. [55] Shop floor management 1. Used the Kanban system.
2. Attached RFID tags to operators, workstations, containers, and inventories.
Cheng et al. [15] Shop floor control and management 1. Collected information on, e.g., machine availability, WIP level.
Zhang et al. [54] Scheduling 1. Collected information on machine use and current status.
Zhang et al. [58] Scheduling 1. Collected information on the progress of items on an unpaced assembly line.
Zhang et al. [59] Logistics 1. Collected information on the operator, transported item, and location of a trolley.

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
4 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Table 2
Applications of Kinect to manufacturing.

Reference Field Applications

Du et al. [18] Factory automation 1. An industrial robot imitated the motion of an operator.
Lisboa et al. [31] Factory automation 1. An industrial robot imitated the motion of an operator.
2. Voice commands were used to control other functions.
Gil et al. [24] Factory automation 1. An industrial robot imitated the motion of an operator.
Chen et al. [14] Computer-aided manufacturing 1. The gesture and motion of an operator were mapped to a CNC operation.
Gil et al. [24] Factory automation 1. An industrial robot imitated the motion of an operator.
2. A seven-step recognition procedure was established.

industrial control system by modifying the codes running in the as a cloud service, and running simulations as a cloud service. A
programmable logic controller, thus revealing the high risk of user is allowed to directly communicate with the cloud via prag-
exposing a machine to the outside environment. matics tools [23].
Both GPS and GSM have been extensively applied to trace the
location and progress of a delivered order [34], whereas WiFi can 3.3. CM and internet of things
be applied for ubiquitous positioning, diagnoses, and control in-
side a factory [3]. For example, a WiFi-based wireless networked CM places virtualized and encapsulated manufacturing re-
control system (WNCS) was established in Refaat et al. [38] for sources in clouds to enable them to be ubiquitously accessed [50].
workcell concatenation. Wireless standards and signal inter- Unlike in distributed manufacturing, CM users do not need to
ference are two concerns of these applications. For example, know where the manufacturing clouds are located. CM can be
Anand et al. [2] compared the performances of two WNCSs based considered a special type of UM that requires network im-
on IEEE 802.11a, b, g (WiFi) and IEEE 802.15.1–2005 (Bluetooth plementation. A recent survey of CM referred to the study by Ren
v1.2), respectively. In Cheng et al. [15], a shop floor supervisor used et al. [39]. Some advances in CM are reviewed as follows.
a handheld device to access information on machine availability, There are two versions of CM: (1) the direct application of
WIP level, and job status transmitted via GSM or WiFi when he or cloud computing techniques to manufacturing and (2) manu-
she was on the move. facturing-oriented cloud computing [51]. Wang and Xu [47] es-
tablished a CM system in which a client interface was provided for
3.2. Web services a user to input a service request. This request was encoded in
extensible markup language (XML) to enable interoperability. A
Computer-aided design and manufacturing software vendors series of tasks were planned by the system server to deliver the
such as AutoCAD have provided Web-based services for product service request. The tasks were mapped to suitable manufacturing
designers to use any place, at any time, and using any device [4]. In services provided by partners worldwide. Manufacturing data are
addition, MISs such as customer relationship management and typically generated according to the ISO 10303 (STEP) standard to
enterprise resource planning systems have been moved onto the ensure integrity; however, this is not conducive to integrating
Web to provide factories with ubiquitous accesses to these sys- manufacturing data that are being transmitted through a dis-
tems [16,20]. tributed manufacturing network. A new capability added to the
Factories establishing websites or online systems for customers STEP standard for supporting XML data structures can provide a
to order or query the status of an order, the so-called business-to- solution to this problem. From this perspective, Valilai and
customer or business-to-business e-commerce, is also an appli- Houshmand [44] established a CM platform to support colla-
cation of UM, because these systems can be accessed from any- boration among manufacturers and the integration of manu-
where at any time. An example is TSMC-Online, which was es- facturing according to the STEP standard. Chen et al. [13] estab-
tablished by Taiwan Semiconductor Manufacturing Company, the lished a CM system for simulating a factory online by using cloud
largest wafer foundry in the world [42]. services. The system imported factory and simulation scenario
Chen and Lin [12] established a digital equipment identifier data in XML format through a CSP and generated an output report
(DEI) system that assigned each piece of equipment a unique that was also in XML format for export to the CSP. Kendrick et al.
identifier, according to which various applications could be prac- [27] proposed the concept of a hybrid manufacturing system, de-
ticed. Using the DEI, a user could gain ubiquitous access to basic fined as a manufacturing system that employs a plurality of pro-
data and up-to-date information of the machine through the DEI cesses on a single piece of equipment. Hybrid manufacturing
system. systems were believed to be an effective unit of decentralized
Web services support predictive maintenance, in which the manufacturing networks, such as a CM or UM network. According
status of a machine is monitored using sensors and the results are to Schlechtendahl et al. [40], the control function of a manu-
analyzed by Web services following the procedures of registering, facturing system can be placed in a cloud and then connected to
deploying, finding, and binding [60,8]. The predictive maintenance the drivers, sensors, and devices of machines. Thus, the control
of a machine can then be arranged to prevent potential break- system becomes scalable and can process highly complex com-
downs. This reduces the time and costs required for vendor putational tasks while retaining the manufacturing information.
technicians to check and repair the machine onsite, thereby pro- To simulate a factory online by using cloud services, the simulation
viding a remote and ubiquitous diagnosis capability. However, to burden must be shared among clouds according to a precise es-
perform an effective analysis, a Web service should be equipped timation of the required simulation time. To estimate this, Chen
with some kind of intelligence, which has yet to be developed. and Lin [12] proposed a fuzzy collaborative method. Mapping a
Chen [10] identified six cloud services that could assist in the service request to possible cloud services is a critical task in CM.
simulation of a factory online: statistical analysis as a cloud ser- For this purpose, Zhang et al. [60] established an intelligent system
vice, equipment definitions as a cloud service, facility layout as a by using the gray relational analysis method. 3D printing, or ad-
cloud service, defining recipes as a cloud service, recipe conversion ditive manufacturing, is another advance in manufacturing that

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 5

contributes to realizing Industry 4.0. Numerous clouds provide 4. Problems and challenges
additive manufacturing services [7]. However, such services are
still limited to products that require high customization and do not 4.1. Problems and challenges
include products that can be mass produced. An Internet of Things
network can also be formed by connecting 3D printers to provide According to Putnik et al. [35], the successful implementation
ubiquitous additive manufacturing services. of UM lacked organizational methods (including mediation and
brokerage services) and information and communication tech-
3.4. Real-time decision making nology infrastructures. The problems and challenges faced by a
factory when establishing a UM system include the following:
UM entails product manufacturing at or delivery to places near
a customer. For this purpose, the availability and capabilities of (1) Reluctance to adopt UM technologies: Social awareness and
nearby manufacturing facilities are considered. However, manu- acceptance of UM is inadequate.
(2) Legal issues: Transmitting data via wireless networks or up-
facturing facility availability changes rapidly. In addition, the cus-
loading data onto a Web or cloud service may result in acci-
tomer may relocate. Therefore, real-time information on customer
dental or deliberate data disclosure. In addition, the upload
location and the availability of each nearby manufacturing facility
and download of a 3D model should be protected by copyright
is critical to the effectiveness of UM. RFID and a GPS are two of the
law, which differs from country to country. To transfer the
most effective means of collecting such real-time data [34,54,9].
processing of an order to another factory with available ca-
Embedding sensors into machines and periodically checking the
pacity, the recipe may need to be converted, adding another
readings helps in correcting minor problems in real time without
risk of leaking technical secrets. Data privacy, confidentiality,
the aid of an external technician [26].
and security is a major concern among most manufacturers.
For the same reason, the reasoning mechanism must be suffi-
(3) Knowledge and trust management: Converting the recipe for
ciently quick to act before the two conditions change. For example,
manufacturing a product in one factory to make the product
in a study by Chen [11], to find a nearby service that could mini- able to be manufactured in another factory is essential to the
mize customer waiting time, a fuzzy mixed-integer programming interoperability and scalability of both factories and relies on
problem needed to be solved; this was a time-consuming process. their mutual trust. Managing the knowledge behind a 3D
To increase problem-solving speed, a fuzzy collaborative problem- model is a challenging task for a ubiquitous additive manu-
solving strategy was proposed. Zhang et al. [59] applied an analytic facturing network.
hierarchy process (AHP) to determine the weights of the attributes (4) The integration of heterogeneous manufacturing resources,
of a manufacturing service. To process a request, a composition of including the encapsulation of a manufacturing resource and
some manufacturing services was formed to maximize the virtualization of a manufacturing service: In practice, different
weighted sum of the reciprocal of the distance, priority, and used sensors are used to trace operators, raw materials, and pallets.
capacity. A similar approach was employed by Zhang et al. [57], The problem is how to wrap these smart devices so that they
through which six attributes were considered, namely cost, lead can be applied to a new manufacturing process in a plug-and-
time, compatibility, on-time delivery probability, reliability, and play manner [52-56,60].
energy. The performance of a manufacturing service regarding the (5) The huge costs for ubiquitously deploying manufacturing re-
six attributes was normalized and subsequently employed to cal- sources [10,17,23,35,52-56].
culate the relational coefficient matrix. This matrix was then (6) A factory may be reluctant to wait (or pay) for maintenance
multiplied to the weight matrix (determined using AHP) to de- service from an overseas equipment vendor, and resorts to a
termine the overall performance of the manufacturing service. local business instead [26]. A remote diagnostic system helps
Overall, applying the calculations that were used under these prevent this, and makes a vendor technician more prepared.
methods to real-time decision making was simple.
4.2. Possible solutions
3.5. Social networks
The following solutions have been or can be used to solve these
According to Foust [21], the relationships among facilities in a problems:
UM system approximate a perfect competition, in which the
profit-maximizing strategy is to produce at the least-cost location. (1) Zhang et al. [52-56] established a smart object gateway
Lyytinen and Yoo [32] asserted that UC (and UM) was going to system, in which a server computer was equipped with several
transform business organizations. More and more service agree- communication interfaces connected to various types of sen-
ments are expected to be signed between companies to colla- sors and smart objects. Thus, heterogeneous manufacturing
boratively achieve UM [26]. Tao et al. [43] observed that the resources could be integrated.
sharing of manufacturing resources and capabilities is moving (2) To the same purpose, the cloudlet2 architecture dashboard
toward socialization. system established in Ferreira et al. [23] read and summarized
Self-organizing networks of autonomous sensors and actuators data from the databases of heterogeneous information
are expected to increase the flexibility, reconfiguration, and fault systems.
tolerance of a UM system [35]. When the available capacity of a (3) Bornhövd et al. [5] established an auto-ID infrastructure that
machine is encapsulated as a manufacturing service and put on a used mapping rules to convert RFID or auto-ID data to
cloud, it can be aggregated with that of a same machine from information that could be fed into an MIS. Thus, data from
heterogeneous sensors and systems could be aggregated.
another factory. A social network, called a virtual capacity net-
(4) In the DEI system [12], each machine was assigned a unique
work, can then be formed by gathering the factories that intend to
DEI code on the basis of the number of applications that could
acquire or rent the machine with the factories that provide the
be developed. For example, it became easier to build a factory
machine for rent or sale, along with information service providers
that act as intermediaries to match the demand for and supply of
the machine [12]. 2
A cloudlet is an information system residing in a cloud.

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
6 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎

simulation model by accessing the basic data and pictures of Acknowledgments

machines through the DEI system. According to the DEI of a
machine, a hub gathering the demand and supply of the This work was supported by the Ministry of Science and
machine could also be established to facilitate the sharing of Technology of Taiwan under grant MOST 103-2221-E-035-073-
unused capacity, thereby increasing the possibility that a MY3.
product was manufactured in more places.
(5) UM does not necessarily mean that a product is present or
producible at every geographical point of a region or country. References
The relaxed version of UM features an environment in which a
product is so extensively available that whenever a demand [1] G. Alexandersson, The Industrial Structure of American Cities, The University
appears, opportunities exist for producing it in the vicinity of Nebraska Press, Lincoln, 1965.
[21]. Additive manufacturing helps to achieve this through the [2] D.M. Anand, J.R. Moyne, D.M. Tilbury, Performance evaluation of wireless
networks for factory automation applications, in: Proceedings of IEEE Inter-
increasing use of 3D printers; thus, an order can be easily national Conference on Automation Science Engineering, 2009, pp. 340–346.
transmitted to a location near a customer, resulting in a [3] J. J. Astrain, J. Villadangos, J. R. Garitagoitia, J. R. G. de Mendívil, V. Cholvi, Fuzzy
ubiquitous additive manufacturing network. location and tracking on wireless networks, in: Proceedings of the 4th ACM
International Workshop on Mobility Management and Wireless Access, 2006,
pp. 84–91.
[4] AutoCAD What is AutocAD 360? 〈https://www.autocad360.com/what-is-auto
cad-360/〉, 2015.
[5] C. Bornhövd, T. Lin, S. Haller, and J. Schaper, Integrating automatic data ac-
5. Discussion and conclusions
quisition with business processes experiences with SAP's auto-ID infra-
structure, in: Proceedings of the Thirtieth International Conference on Very
There is always demand for UM. For example, customers are Large Databases, 2004, pp. 1182–1188.
[6] I. Bose, R. Pal, Auto-ID: managing anything, anywhere, anytime in the supply
willing to pay a higher price to get a new iPhone 6s sooner. Users
chain, Commun. ACM 48 (8) (2005) 100–106.
are also willing to pay a higher fee to get their iPhones repaired at [7] A. Brant, M.M. Sundaram, A novel system for cloud-based micro additive
closer locations. Such demand, if not satisfied, becomes opportu- manufacturing of metal structures, J. Manuf. Process. 20 (3) (2015) 478–484.
[8] S.K. Cha, J.Y. Song, J.S. Choi, u-Manufacturing needs M2M (machine to ma-
nities for others. chine) for ubiquitous computing world, in: Proceedings of the International
UM can be roughly divided into the applications of ubiquitous Conference on Control, Automation and Systems, 2005, pp. 1–3.
technologies and the widespread deployment of manufacturing [9] T. Chen, A fuzzy parallel processing scheme for enhancing the effectiveness of
a dynamic just-in-time location-aware service system, Entropy 16 (2014)
facilities. In the past, most UM applications belonged to the first 2001–2022.
type; they were confined to in-factory operations or logistics. [10] T. Chen, Strengthening the competitiveness and sustainability of a semi-
conductor manufacturer with cloud manufacturing, Sustainability 6 (2014)
Applications for the second type were yet to be developed. With 251–268.
the advances in additive manufacturing, CM, cloud computing, [11] T. Chen, Enhancing the performance of a ubiquitous location-aware service
system using a fuzzy collaborative problem solving strategy, Comput. Ind. Eng.
CPS, and Web services, increasing amounts of manufacturing re-
87 (2015) 296–307.
sources, services, and knowledge are moving onto the Internet to [12] T. Chen, Y.-C. Lin, A digital equipment identifier system, J. Intell. Manuf. (2015),
be shared, which grants factories ubiquitous access to such re- in press.
[13] T. Chen, Y.-C. Wang, A. Al-Refaie, Establishing a factory simulation system for
sources, services, and knowledge. As the capabilities of factories cloud-based simulation services, Int. J. Internet Manuf. Serv. (2015), in press.
are enhanced, we can expect that products can be manufactured [14] T. Chen, Y.-C. Wang, Z. Lin, Predictive distant operation and virtual control of
computer numerical control machines, J. Intell. Manuf. (2015), in press.
more ubiquitously, yielding applications of the second type.
[15] M. Cheng, R. Y. Zhong, Y. Li, H. Luo, S. Lan, G. Q. Huang, Cloud service-oriented
In addition, except for a few cases, most UM applications have dashboard for work cell management in RFID-enabled ubiquitous manu-
never undergone cost–benefit analyses. One reason for this is the facturing, in: Proceedings of the 10th IEEE International Conference on Net-
working, Sensing and Control, 2013, pp. 379–382.
difficulty of collecting cost and benefit information. However, for a [16] Cloud Sherpas, 〈http://www.cloudsherpas.com/cloud-services/manufacturing/
UM technology to be widely applied, a credible cost–benefit ana- 〉, 2013.
[17] M. Davidson, Why cloud manufacturing software adoption rates are rising,
lysis must be conducted.
〈https://www.manufacturing.net/articles/2013/10/why-cloud-manufacturing-
As Industry 4.03 [22] is adopted across the world, how UM can software-adoption-rates-are-rising〉, 2013.
contribute becomes a critical challenge. As we observed, it ap- [18] G. Du, P. Zhang, J. Mai, Z. Li, Markerless Kinect-based hand tracking for robot
teleoperation, Int. J. Adv. Robot. Syst. 9 (2012) 36.
peared that Industry 4.0 had placed an undue emphasis on CPS. [19] L. Elliott, The business of ICT in manufacturing in Africa, 〈http://www.afribiz.
When a machine is connected to the Internet, it can be made info/content/2010/the-business-of-ict-in-manufacturing-in-africa〉, 2010.
identifiable by the DEI system according to which factories using [20] Epicor, End-to-end on-demand ERP solution for manufacturers, 〈http://www.
epicor.com/Cloud/Pages/Cloud-Manufacturing.aspx〉, 2013.
the machine form a social network. The recipes for manufacturing [21] B.J. Foust, Ubiquitous manufacturing, Ann. Assoc. Am. Geogr. 65 (1) (1975)
a product can then be converted from factory to factory by using 13–17.
[22] Federal Ministry of Education and Research, Project of the future: Industry 4.0.
Web services, thereby enabling the manufacturing of the product 〈http://www.bmbf.de/en/19955.php〉, 2013.
across factories (i.e., ubiquitously). Accordingly, UM becomes a [23] L. Ferreira, G. Putnik, M. Cunha, Z. Putnik, H. Castro, C. Alves, V. Shah, M.L.
R. Varela, Cloudlet architecture for dashboard in cloud and ubiquitous man-
realizable target for Industry 4.0, even though Industry 4.0 covers
ufacturing, Procedia CIRP 12 (2013) 366–371.
a greater number of more diverse fields than UM does. [24] P. Gil, C. Mateo, F. Torres, 3D visual sensing of the human hand for the remote
The success of UM obviously depends on the quality of the operation of a robotic hand, International Journal of Advanced Robotic Sys-
tems 11 (26) (2014).
manufacturing services deployed. For example, many Japanese [25] G.Q. Huang, P.K. Wright, S.T. Newman, Wireless manufacturing: a literature
manufacturers have deployed manufacturing facilities globally review, recent developments, and case studies, Int. J. Comput. Integr. Manuf.
such as in China and Malaysia. However, to the people in these 22 (7) (2009) 579–594.
[26] K. Jonsson, U.H. Westergren, J. Holmström, Technologies for value creation: an
countries, a product “made in Japan” is still favorable to a domestic exploration of remote diagnostics systems in the manufacturing industry, Inf.
product, even though the local manufacturing facilities are the Syst. J. 18 (3) (2008) 227–245.
[27] B.A. Kendrick, V. Dhokia, S.T. Newman, Strategies to realize decentralized
actual producers. manufacture through hybrid manufacturing platforms, Robot. Comput. Integr.
Manuf. (2015), in press.
[28] S.K. Khaitan, J.D. McCalley, Design techniques and applications of cyberphy-
3
Industry 4.0 is the introduction of the Internet of Things and Services into the sical systems: a survey, IEEE Syst. Mag. 9 (2) (2014) 350–365.
manufacturing environment [22]. [29] S. Karnouskos, Stuxnet worm impact on industrial cyber-physical system

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i
 T. Chen, H.-R. Tsai / Robotics and Computer-Integrated Manufacturing ∎ (∎∎∎∎) ∎∎∎–∎∎∎ 7

security, in: Proceedings of the 37th Annual Conference on IEEE Industrial cloud computing paradigm, Robot. Comput. Integr. Manuf. 29 (1) (2013)
Electronics Society, 2011, pp. 4490–4494. 110–127.
[30] J. Lambrecht, M. Kleinsorge, M. Rosenstrauch, J. Krüger, Spatial programming [45] Virtual Reality Society, Virtual reality, 〈http://www.vrs.org.uk/〉, 2009.
for industrial robots through task demonstration, Int. J. Adv. Robot. Syst. 10 [46] J. Wang, L. Zhang, L. Duan, R.X. Gao, A new paradigm of cloud-based predictive
(2013) 254. maintenance for intelligent manufacturing, J. Intell. Manuf. (2015), in press.
[31] H. B. Lisboa, L. A. R. de Oliveira Santos, E. R. Miyashiro, K. J. Sugawara, P. E. [47] X.V. Wang, X.W. Xu, An interoperable solution for Cloud manufacturing, Robot.
Miyagi, F. Junqueira, 3D virtual environments for manufacturing automation, Comput. Integr. Manuf. 29 (4) (2013) 232–247.
in: Proceedings of the 22nd International Congress of Mechanical [48] M. Weiser, Ubiquitous computing, 〈http://www.ubiq.com/hypertext/weiser/
Engineering, 2013. UbiHome.html〉, 1988.
[32] K. Lyytinen, Y. Yoo, Issues and challenges in ubiquitous computing, Commun. [49] K.H.M. Wong, P.C.L. Hui, A.C.K. Chan, Cryptography and authentication on RFID
ACM 45 (2002) 63–65. passive tags for apparel products, Comput. Ind. 57 (4) (2006) 342–349.
[33] A. Pham, E3: Microsoft shows off gesture control technology for Xbox 360, Los [50] D. Wu, J.L. Thames, D.W. Rosen, D. Shcaefer, Enhancing the product realization
Angeles Times, 2009. process with cloud-based design and manufacturing systems, J. Comput. Inf.
[34] K.R. Prasanna, M. Hemalatha, RFID GPS and GSM based logistics vehicle load Sci. Eng. 13 (2013) 041004-1–041004-14.
balancing and tracking mechanism, Procedia Eng. 30 (2012) 726–729. [51] X. Xu, From cloud computing to cloud manufacturing, Robot. Comput. -Integr.
[35] G. Putnik, C. Cardeira, P. Leitão, F. Restivo, J. Santos, A. Sluga, P. Butala, Towards Manuf. 28 (1) (2012) 75–86.
ubiquitous production systems and enterprises, in: Proceedings of IEEE In- [52] Y. Zhang, G.Q. Huang, T. Qu, O. Ho, S. Sun, Agent-based smart objects man-
ternational Symposium on Industrial Electronics, 2007, pp. 3203–3208. agement system for real-time ubiquitous manufacturing, Robot. Comput.
[36] G. Putnik, Advanced manufacturing systems and enterprises: Cloud and ubi- -Integr. Manuf. 27 (3) (2011) 538–549.
quitous manufacturing and an architecture, J. Appl. Eng. Sci. 10 (3) (2012) [53] Y. Zhang, G.Q. Huang, T. Qu, S. Sun, Real-Time Work-in-progress Management
127–143. For Ubiquitous Manufacturing Environment. Cloud Manufacturing, Springer,
[37] G. Putnik, Z. Putnik, A semiotic framework for manufacturing systems in- London (2013), pp. 193–216.
tegration–part I: generative integration model, Int. J. Comput. Integr. Manuf. [54] Y. Zhang, G.Q. Huang, S. Sun, T. Yang, Multi-agent based real-time production
23 (8–9) (2010) 691–709. scheduling method for radio frequency identification enabled ubiquitous
[38] T. K. Refaat, R. M. Daoud, H. H. Amer, M. Hassan, O. M. Sultan, Workcell con- shopfloor environment, Comput. Ind. Eng. 76 (2014) 89–97.
catenation using WiFi-based wireless networked control systems, in: Pro- [55] Y. Zhang, P. Jiang, G. Huang, RFID-based smart kanbans for just-in-time
ceedings of the 17th IEEE International Conference on Electronics, Circuits, manufacturing, Int. J. Mater. Prod. Technol. 33 (1–2) (2008) 170–184.
and Systems, 2010, pp. 138–141. [56] Y.F. Zhang, T. Qu, O. Ho, G.Q. Huang, Real-time work-in-progress management
[39] L. Ren, L. Zhang, F. Tao, C. Zhao, X. Chai, X. Zhao, Cloud manufacturing: from for smart object enabled ubiquitous shop floor environment, Int. J. Comput.
concept to practice. Enterprise, Inf. Syst. 9 (2) (2015) 186–209. Integr. Manuf. 24 (5) (2011) 431–445.
[40] J. Schlechtendahl, F. Kretschmer, Z. Sang, A. Lechler, X. Xu, Extended study of [57] Y. Zhang, D. Xi, R. Li, S. Sun, Task-driven manufacturing cloud service proactive
network capability for cloud based control systems, Robot. Comput. Integr. discovery and optimal configuration method., Int. J. Adv. Manuf. Technol.
Manuf. (2015), in press. (2015), in press.
[41] M. Strassner, T. Schoch, Today's impact of ubiquitous computing on business [58] Y. Zhang, J. Xu, S. Sun, T. Yang, Real-time information driven intelligent navi-
processes, in: Proceedings of the First International Conference on Pervasive gation method of assembly station in unpaced lines, Comput. Ind. Eng. 84
Computing, 2002. (2016) 91–100.
[42] Taiwan Semiconductor Manufacturing Company, TSMC-Online. 〈http://online. [59] Y. Zhang, G. Zhang, W. Du, J. Wang, E. Ali, S. Sun, An optimization method for
tsmc.com/online/〉, 2015. shopfloor material handling based on real-time and multi-source manu-
[43] F. Tao, Y. Cheng, L. Zhang, A.Y.C. Nee, Advanced manufacturing systems: so- facturing data, Int. J. Prod. Econ. 165 (2015) 282–292.
cialization characteristics and trends, J. Intell. Manuf. (2015), in press. [60] Y. Zhang, G. Zhang, Y. Liu, D. Hu, Research on services encapsulation and
[44] O.F. Valilai, M. Houshmand, A collaborative and integrated platform to support virtualization access model of machine for cloud manufacturing, J. Intell.
distributed manufacturing system using a service-oriented approach based on Manuf. (2015), in press.

Please cite this article as: T. Chen, H.-R. Tsai, Ubiquitous manufacturing: Current practices, challenges, and opportunities, Robotics and
Computer Integrated Manufacturing (2016), http://dx.doi.org/10.1016/j.rcim.2016.01.001i