Design of Driver Assist
Autonomous Navigation System
For BAE Systems
Team 1
Joe Dubovi
Seth Forney
Rahul Krishna
Tim Zetterberg
EDSGN 100
Penn State
April 25, 2008
Design of Driver Assist
Autonomous Navigation System
For BAE Systems
Team 1
Joe Dubovi
Seth Forney
Rahul Krishna
Tim Zetterberg
EDSGN 100
Penn State
April 25, 2008
Executive Summary
This report presents an overview of the Driver Assist Autonomous Navigation System.
This system is designed for military vehicles to assist in navigation during blackout conditions.
Thus, this system is able to detect and avoid obstacles while keeping three vehicles in formation.
This was accomplished by defining the needs for the system based on the given requirements, by
conducting research into various sensor systems, and by constructing selection matrices.
These processes determined that the best design to pursue was a combination of Adaptive
Cruise Control (ACC), Global Positioning System (GPS), Light Detection and Ranging
(LIDAR), and Radio Detection and Ranging (RADAR). The ACC is mounted on the front of the
vehicle and allows the vehicles to stay in formation, the GPS receiver is inside the vehicle and
allows for navigation in a broad scope, the LIDAR is mounted above the vehicle and is able to
provide a very large amount of information about a medium-range distance (about fifty meters),
and the RADAR is mounted above the vehicle and is able to provide substantial information
about a relatively long distance (about 100 meters). This concept is further defined in the form of
a prototype and Computer Aided Design (CAD) models.
Introduction
This system is designed for use on military vehicles to increase driver safety. Often,
convoys must maintain close spacing on damaged road surfaces in blackout conditions.
Therefore, a system that is capable of warning drivers of upcoming hazards can be of extreme
importance to their survival. However, sensor systems can be very expensive. Because of this, it
was necessary to find a reasonable balance between the information provided by the system and
the price of the system
Needs-Metrics Matrix
Based upon the information provided in the project description, a list of needs to be met
by the system was created. A list of metrics was then created based upon calculations performed
regarding the capabilities of the vehicle to be used. After determining the turning radius of the
vehicle using the equation a=v2/r, trigonometry and the arc-length formula were used to
determine how far the car would travel before being out of the way of an approaching stationary
obstacle. Through this, it was determined that a stationary obstacle would have to be detected, at
the very latest, when the vehicle is 25 meters away from the vehicle. Thus, to allow for increased
safety, it was decided that the system should detect stationary obstacles within 50 meters. A
similar calculation was performed that took the speed of an oncoming vehicle into account. This
calculation determined that an oncoming vehicle should be detected within at least 75 meters.
These needs and metrics were then compiled into a needs-metrics matrix (Table 1).
Table 1. Needs-Metrics Matrix: Each row corresponds to a different need. Each column corresponds to a
different metric. An bullet denotes which need is met by which metric.
Position of vehicle/obstacle
Can suggest an avoidance
vehicles up to 75 m away
Detects stationary object
Can alert the driver of an
Provides information for
driver to keep 5m gap in
is continually detected
obstacle within 1 ms
strategy within 2 ms
Map the position of
every half a second
within 50 meters
convoy
Detects obstacle •
Suggested avoidance • •
Locate obstacle •
Detect vehicle in convoy •
Locate vehicles • •
Stay in formation •
Locate oncoming vehicles •
Concept Screening
Using the needs-metrics matrix as a guide, research was conducted into various sensor
types in order to determine what sensors would be plausible options to meet the necessary needs.
With the knowledge gained by the team, each member individually created a list of ten possible
ways to solve the problem. After removing duplicates and those deemed completely implausible
(i.e. animals, etc), the twenty five remaining concepts were placed in a screening matrix (Table
2) to determine the best options. The complete list of needs can be found in Appendix A. These
concepts were evaluated based on ten criteria developed from the listed needs and the project
description. The seven highest scoring concepts (highlighted in the matrix) continued to the
scoring phase.
Table 2. Concept Screening Matrix: The numbers across the top represent concepts. The needs are found
on the leftmost column of the matrix. A plus or minus sign represents whether the above concept was
determined to be respectively better or worse than the baseline.
1 2 3 4 5 6 7 8 9 10 11 12 13
Cost - + + 0 0 + - + 0 - 0 - +
Space Efficiency 0 + + 0 0 - - + 0 0 0 - +
Durability 0 + + 0 0 - - + 0 0 0 - +
Ease of Use + - - 0 0 + - - 0 0 - - -
Reliability + 0 0 0 0 - + - - - 0 - 0
Feasability 0 0 0 - 0 + - - - - 0 - 0
Accuracy + 0 + + + + + + + + + + +
Safety + 0 + - + - 0 - - + + + 0
Maintainability 0 0 0 0 0 - - 0 0 0 0 - 0
Range - 0 0 - 0 - - - - - 0 - -
Sum 2 2 4 -2 2 -2 -5 -1 -3 -2 1 -6
14 15 16 17 18 19 20 21 22 23 24 25
Cost 0 - + + + + - + + + - 0
Space Efficiency 0 0 + + + + - - + - - 0
Durability 0 0 + + + + - - - + - 0
Ease of Use 0 - - - - - + + - + - 0
Reliability + - - - - - - - - - - 0
Feasability 0 - 0 0 0 - + + - + - 0
Accuracy + + 0 + + + + + - - 0 0
Safety 0 + 0 + + - 0 - - - - 0
Maintainability 0 0 0 0 0 0 0 - - + - 0
Range - - 0 - 0 - - - - - - 0
Sum 1 -3 1 2 3 -1 -2 -2 -6 0 -9 0
Concept Scoring
Before being placed into a scoring matrix (Table 3), the seven best concepts were
reviewed and combined to be more easily evaluated. Specifically, it was determined that every
major concept should have both Adaptive Cruise Control (in order to best meet the need of
convoy formation) and GPS (in order to have broad navigation capabilities). These features are
also both inexpensive, so it was assumed that they would not have a large affect on the cost
aspect of the project. Also, they did not warrant being their own concept in the matrix, as they
surely would not be able to meet the needs of the project on their own. These new combined
concepts were placed into a scoring matrix and evaluated based upon similar criteria to the
screening matrix. In the scoring matrix, each criterion was assigned a weight, based on how
important each was determined to be. The sum of the weights is 100, as that would be the total if
each criterion had a value of ten. A baseline was selected for each criterion, with the goal being
that the baseline was the concept that was most “average” in the criterion. The baselines were
assigned a value of three. Each concept was then judged against the baseline on a scale of 1-5, 1
being much worse than the baseline, 5 being much better than the baseline, and 3 being equal to
the baseline. Each of those values was then multiplied by the weight and totaled to determine the
best concept.
Concept A
The first concept uses LIDAR to map the terrain ahead and any oncoming obstacles. It is
able to provide a very large amount of information with a range of about 50 meters. The LIDAR
is mounted on the top of the vehicle. In addition, it has ACC to stay in formation and GPS for
broad navigation.
Concept B
The second concept uses SONAR to sense oncoming
obstacles. However, it is able to provide information at only a
relatively short range. The SONAR is mounted on top of the vehicle.
In addition, it also has ACC to stay in formation and GPS for broad
navigation.
Concept C
The third concept uses RADAR to detect obstacles. It can
provide an adequate amount of information with a range of about 100
meters. The RADAR is mounted above the vehicle. As with the other
concepts, it also has ACC to stay in formation and GPS for broad
navigation.
Concept D
The fourth concept uses a LASER Range Finder to detect
obstacles. It can provide information at a mid-range distance. The
LASSER Range Finder is mounted above the vehicle. In addition,
this concept also has ACC to stay in formation and GPS for broad
navigation.
LIDAR with ACC and GPS emerged as the winner. However, RADAR and SONAR
were very close. The LASER Range Finder emerged as the clear loser. It should be noted that it
was later discovered that LASER Range Finders are very similar to LIDAR, but the LASER
Range Finder was judged based solely under information discovered about products specifically
called LASER Range Finders. The information used to score the LIDAR was based on the
Velodyne HDL-64E LIDAR (Velodyne). The information used to score the SONAR (acoustic
microphone) was obtained from the official project presentation created by BAE Systems. The
information used to score RADAR was based on the Smartmicro Stop & Go Single Sensor
System (Smartmicro). The information used to score the Laser Range Finder was based upon
various similar existing products.
It was decided that because of the close scores in the matrix, the final system would have
both LIDAR and RADAR. The LIDAR is able to provide a large amount of detail at mid-range
distances (about 50m), while the RADAR is able to provide a smaller amount of detail at longer
distances (at least 100m). The LIDAR also acts as something of a backup, as RADAR is not
difficult to jam (by transmitting on the same frequency as is used by the RADAR). The SONAR
was deemed unnecessary, as it obtained the smallest score, and would be superfluous when
combined with the LIDAR and the RADAR.
Table 3. Concept Scoring Matrix: The criteria are listed on the left; the concepts are listed across the top.
Underlined numbers represent baselines.
Weight A B C D
Cost 5 2 3 5 4
Space Efficiency 5 3 4 2 3
Durability 10 2 4 3 2
Power
Consumption 10 3 4 2 3
Detection Time 15 3 1 4 2
Feasability 5 3 1 3 2
Accuracy 15 5 4 3 2
Weight 10 3 4 2 2
Maintainability 10 2 3 3 2
Range 15 3 2 3 3
Sum 305 295 300 240
A: LIDAR, ACC, GPS
B: SONAR, ACC, GPS
C: RADAR, ACC, GPS
D: LASER Range Finder, ACC,
GPS
Cost Analysis
Using the costing guidelines provided by BAE Systems, an estimate of the cost was
determined. Based upon information obtained from researching various similar sensors on the
internet, it was estimated that the cost of LIDAR is about $75,000, the cost of RADAR is about
$50,000, the cost of ACC is about $2,000, and the cost of GPS is about $1,000. It was also
assumed that the entire system would take two workers (paid $30/hr) about 100 hours per vehicle
to install. Also, as the guidelines stated that testing would total $2880, and that the total cost
would equal 1.3 times the base cost. This brings the total cost of the system to $526,344.
Description of Final Concept
The final concept includes a single LIDAR unit, a single RADAR unit, a single ACC unit
(which works by RADAR, but will be devoted specifically to convoying), and a GPS unit. The
LIDAR is mounted on the roof of the vehicle, and serves to map the terrain ahead of the vehicle
to a range of about 50 meters. This information can be used by the driver to determine what parts
of the road are safe. The RADAR is mounted above the vehicle, and is used to look farther ahead
(about 100 meters) to detect oncoming obstacles and vehicles in enough time to swerve to avoid
them. It should be noted that RADAR is not difficult to jam, in which case the LIDAR can also
be used to detect obstacles. The ACC unit is mounted on the front of the vehicle, and is used to
keep a preset distance behind the leading vehicle. It can also automatically adjust its speed if the
leading vehicle does so, allowing the driver to focus on other hazards. Finally, the GPS receiver
is located inside the vehicle, and is simply used for broad navigation. It is not technically part of
the hazard avoidance, but it seems necessary to include nonetheless.
All of these sensors feed into a display located inside the vehicle. This display is
completely waterproofed in case of vehicle damage. The LIDAR maps the road, while the
RADAR alerts the driver to oncoming obstacles. The vehicle does not provide any avoidance
measures of its own except for the ACC, which will automatically adjust the speed of the vehicle
to keep a uniform distance behind the vehicle in front of it. Instead, the display provides enough
information for the driver to make his/her own decision about how to best avoid the obstacle.
Consider a convoy conducting a mission in blackout conditions. The convoy is trying to
avoid detection, while still moving quickly yet safely. The driver watches his display screen,
which presents the terrain as mapped by the LIDAR. Suddenly, the RADAR detects an obstacle
in its range, and displays a warning with the relative position of the oncoming object. The driver
quickly assesses the situation and has enough time to swerve out of the way of the object.
Meanwhile, those following the lead vehicle also get warnings on their screens that the lead
driver has swerved, allowing those following to do the same. The system has enough range that
the driver should be able to swerve to avoid any obstacle, but the choice remains up to the driver
as to whether they want to swerve at full speed or slow down to avoid the object.
Conclusions
This report documents the process through which the Driver Assist Autonomous
Navigation System was designed. The system was to assist the driver of a military vehicle by
providing warning when there are obstacles obstructing the road. Through sensor research, needs
analysis, and decision matrices, a concept was selected to be created in the form of CAD models
and a prototype. The chosen concept was a system that relays information to the driver as
provided by a Light Detection and Ranging unit, a Radio Detection and Ranging unit, an
Adaptive Cruise Control unit, and a Global Positioning System unit.
This system is strong in that it covers its bases. It has both capabilities for detailed short-
range mapping through the LIDAR and long range detection through the RADAR. However, it is
difficult to determine how durable these devices are when actually used in a combat situation. If
this project were to be performed again, it would be helpful to have time with actual sensors in
order to get a better feel for their capabilities. Also, conflicting information from many sources
made finding reliable data difficult. Even so, the project meets all previously set goals, and thus
is a success.
Appendix A. Complete list of screened concepts
1. Lidar, ACC, GPS
2. Sonar, GPS
3. Radar, GPS
4. Video, ACC, GPS
5. Radar, ACC, GPS
6. Long pole detects bump in ground
7. Lidar equipped plane sends info to cars
8. Thermal imaging, GPS
9. Thermal imaging, GPS, ACC
10. Thermal imaging, Lidar, ACC
11. Sonar, Radar, GPS
12. Video, ACC, GPS, Lidar
13. Laser range finder, ACC
14. Laser range finder, ACC, GPS
15. Video, ACC, Lidar
16. Sonar, ACC
17. Lidar, ACC
18. Radar, ACC
19. Thermal imaging, ACC
20. Bulldozer clears way for cars
21. Long pole with laser range finder
22. Trained bird with Lidar
23. Car with front spring to avoid damage in a wreck
24. Car with springs in shocks to propel car into air, Lidar
25. Sonar, GPS, ACC
Red= winners of matrix
References
Smartmicro.<http://smartmicro.de/content/index.php?option=com_content&task=view&id=8&It
emid=6>. 18 April 2008
Velodyne High Definition Lidar. <http://www.velodyne.com/lidar/>. 18 April 2008
