Fully Autonomous Sentry Turret System

ECE 4007 Senior Design Final Project Report

Section L01, Turret Trackers Design Team

Project Advisor, Dr. David Keezer

Patrick Croom
Kevin Neas
Anthony Ogidi
Joleon Pettway

May 5, 2010

Table of Contents
Executive Summary ......................................................................................................... 3
1. Introduction.................................................................................................................. 4
1.1 Objective......................................................................................................... 4
1.2 Motivation...................................................................................................... 5
2. Project Description and Goals ..................................................................................... 6
3. Technical Specification................................................................................................ 7
4. Design Approach and Details
4.1 Design Details .......................................................................................... 9
4.2 Codes and Standards ..................................................................................... 18
5. Schedule, Tasks, and Milestones................................................................................. 18
6. Project Demonstration ................................................................................................ 21
7. Marketing and Cost Analysis ..................................................................................... 24
8. Summary..................................................................................................................... 25
9. References................................................................................................................... 27

Executive Summary
The Turret Trackers design group has successfully developed a fully functional proof-ofconcept prototype Fully Autonomous Sentry Turret system. This system is an autonomous, area
protection, anti-personnel sentry turret with integral Identify Friend/Foe (IFF) capability. The
system prototype uses a standard webcam connected to a notebook computer running custom
software in order to detect and track movement sources within the systems field-of-view. This
data is used to position a pair of servo motors in a dual axis pan and tilt mount in order to aim an
optical laser marker at the moving target. The IFF capability is achieved using two independent
systems. For the first, targets who are attired in the designated safe color scheme will be
allowed to pass unharmed. The second safety uses an IR remote to transmit a safe code to the
turret, allowing safe passage so long as the code keeps transmitting. There is a clear need for
such a system in certain military applications, as no similar product is currently available. The
only other systems capable of fulfilling the same function are of very limited utility and much
more time consuming to set up. Specifically, currently available systems are very ineffective
indoors and are almost entirely indiscriminate in their operation. Both of these failures are
addressed by the FAST system. By using commercial off-the-shelf (COTS) materials as much as
possible, the design team was able to develop the prototype system for only $1893.50.

Fully Autonomous Sentry Turret System

1. Introduction
The Turret Trackers design group has successfully developed a fully functional proof-ofconcept prototype Fully Autonomous Sentry Turret system. This system is an autonomous, area
protection, anti-personnel sentry turret with integral Identify Friend/Foe capability. There is a
clear need for such a system in certain military applications, as no similar product is currently
available. The design and development of the prototype has cost $1893.50.

1.1 Objective
The purpose of this design effort was the development of a viable proof-of-concept
prototype of the Fully Autonomous Sentry Turret (FAST) System. In its simplest form, the
FAST system is composed of a standard web camera connected to a commercial notebook
computer using a custom software package to detect and track sources of movement while
ignoring motion sources designated as friendly. This tracking information will be passed to a
series of servo motors which will align the attached firearm to point and fire at the motion
source. For the prototype, a optical range low power laser was used in place of an actual firearm.
The FAST system is capable of protecting a fixed position from the approach of enemy
combatants from a fixed direction, optimized for use in indoor areas. FAST is also capable of
identifying and not targeting friendlies within its field of vision based on identifying markings.
Authorized personnel can also pass safely through the field-of-fire by transmitting the correct
code with a handheld IR transmitter. In order to accomplish this, FAST employs a computer
vision system to track motion within a fixed field of view and a series of servo motors to orient a
pointing/projectile firing device at the observed motion source. FAST is also able to

differentiate between targets and non-targets based on object size and major identifying
markings. This system is ideally suited for use by military organizations and private security
companies for static defensive purposes. The FAST system has a detection area optimized for
most indoor scenarios and is small enough and light weight enough to be transported and setup
by a single individual.

1.2 Motivation
There exists, among modern military organizations, a distinct lack of available systems
for the purpose of unmanned anti-personnel area denial and/or defense. The only existing
systems, which provide any remotely similar function, are barbed wire and landmines. Both of
these systems are rather dated and generally low-tech. In addition, these existing systems are
entirely indiscriminate in their operation, effectively closing off the position in which they are
emplaced to everyone. These systems are further handicapped by the limitations on the settings
in which they can be employed as well as the time consuming and potentially complicated setup
that is required in order for them to be effective.
The FAST system suffers from none of these inherent limitations. FAST is capable of
discriminating between targets and non-targets in mere fractions of a second. FAST can be
easily emplaced and positioned in less than a minute and can be activated with only a few
keystrokes. This also means that FAST can be employed indoors, something that is generally
impractical for any other system. The FAST system is also superior in many situations to using
human soldiers for the same purpose because FAST cannot become tired or distracted while on
duty. This also means that those same soldiers will be kept out of danger.

2. Project Description and Goals

The FAST system prototype was constructed using readily available COTS materials as
much as possible in order to reduce costs. The external housing of the unit was constructed of
standard plywood supported by 2in. by 4in. lumber beams to create a sturdy cubic container
approximately 2 feet on a side. The back of this box was sealed with a hinged and locked metal
grating door. This entire assemble was set on wheels in order to make it more portable. A pan
and tilt servo motor assembly in a vertical configuration was mounted atop this housing
alongside the primary webcam and IR sensor. Within the housing were placed the control
circuitry for both the servo motors and the IR sensor. The housing also contains the dc power
supply for the servo controller and the standard notebook computer which acts as the primary
processor for the prototype. The prototype is depicted in Figure 1.

Figure 1. External view of the FAST prototype unit.

The FAST system prototype was designed with two primary, overarching goals in mind.
First, can the system successfully engage a target before that target can reach the turret itself?
And second, can a properly designated ally pass through the kill-zone without being engaged?
Both of these primary goals were unequivocally achieved by the prototype system developed by
this design team. On this basis alone, the system prototype can be considered a success.

However, there were some changes to the more technical aspects of the specifications between
the proposed design and the prototype. The design was originally intended to have an effective
detection range of 50 feet and be able to engage targets within 2.0 seconds of detection. For the
prototype, it was necessary to shorten the detection range to 40 feet, but the prototype achieved a
target acquisition time of less than 1.0 seconds. This much faster acquisition time more than
compensated for the reduce range in terms of overall engagement capability.

3. Technical Specifications
The FAST system was designed to meet both quantitative and qualitative specifications.
The quantitative specifications are shown in Figure 1. The design proposed the ability to track
and acquire a target with 50 feet of the device. The webcam processes a resolution of 352x288,
and the IR receiver is reliable up to 45 feet. Due to the limitations of the webcams quality and
the IR receivers range, the FAST system is limited to a 40 feet range. Originally, the field of
view was limited to 45 in order to reduce costs. However, the code acquires targets within a 65
field of view, with the ability to be calibrated for a range up to 90. Finally, in order to feasibly
track and obtain a target as fast as a human, the system required a minimum acquisition rate of 2
seconds. The final design can track movement faster than 15 frames per second. This allows the
servos to acquire the target in less than 1 second and as little as half of a second.
Table 1. FAST System Quantitative Specifications
Feature

Proposed Specifications

Actual Specifications

Effective Range

50 feet

40 feet

Field of View

45 degrees

65+ degrees

Maximum Target Acquisition Rate

2 seconds

< 1 second

The FAST system was required to meet the following qualitative specifications:

Once activated, this system will operate without user interaction

Uses visual monitoring to detect significant sources of movement

Portable

Successfully engage a target before that target can reach the turret itself

Allow properly designated personnel pass through the area-of-defense without being engaged

Identify authorized personnel using a color recognition system

All qualitative specifications were determined to be vital components for the success of the
system. However, security concerns with the color recognition system required that an additional
recognition system be implemented into the system. An infrared security system was
implemented which requires authorized personnel to send a predetermined IR signal
continuously while passing through the area-of-defense. The final qualitative specifications met
are:

Once activated, this system will operate without user interaction

Uses visual monitoring to detect significant sources of movement

Portable

Successfully engage a target before that target can reach the turret itself

Allow properly designated personnel pass through the area-of-defense without being engaged

Identify authorized personnel using a color recognition system

Identify authorized personnel using an IR receiving system

4. Design Approach and Details

4.1 Design Details
In order to be portable and easy to use, the FAST system contained the following in a
2x2 box:

1- Microsoft LifeCam VX-1000 webcam

1- HP laptop

1- 1061 Phidget Advanced Servo 8-Motor control board

2- Hitec HS-645MG Super Torque Servo motors

1- Hitec SPT200 Pan & Tilt System

1- Sony blu-ray player remote

1- Arduino Duemilanove board

1- Panasonic PNA4601M infrared receiver

1- LED

1- 12 volt battery

As shown in Figure 2, the Arduino board and webcam sent data to the notebook through
USB, which in turn sent data to the Phidget control board. The webcam and Arduino board
received power through USB, while the Phidget control board received a 12 volt, 3 amp power
supply from the external battery.

Figure 2. System Configuration

Interface
The FAST system has been design to be user friendly. Once the code is started,
the user will have two options in the File menu: Open Local Capture Device or Exit.
When Open Local Capture Device is selected, the user will be prompted to choose the
webcam, as seen in Figure 3.

Figure 3. System configuration

Most laptops are produced with an internal camera, so this menu allows the user to
choose an external webcam instead. Once the webcam is selected, the FAST system will
continue to run indefinitely without the need for user interaction. While the program runs, the
system will display the webcams video while constantly tracing rectangles around the biggest
area of movement, as seen in Figure 4.

10

Figure 4. Motion Tracking Interface

Motion Tracking
The motion tracking software is implemented using the AForge.Nets computer vision
library. The library is an open source C# framework designed for developers and researchers in
the fields of Computer Vision [1]. The motion tracking class runs under the motion solution
(the code can be accessed through the website) [3]. This solution has been adopted from Andrew
Kirillovs Motion Detection Algorithms software, which has been posted under the GNU
General Public License [2].
In order to begin tracking motion, the code must first connect to a webcam. The interface
will list all image devices connected to the computer. Once the user has selected the appropriate
webcam, the system receives a live digital video feed through the USB. In order to process the
frames at a high frames per second rate (greater than 15 fps), all image comparisons of the video
stream are done in virtual memory. The frames are compared in bitmap form in the computers
memory and are never saved on the hard drive. This allows for faster computation and
comparison of successive frames.
11

Figure 5. Tracking and Targeting Algorithm

The tracking and targeting algorithm can be found in Figure 5. In order to detect
movement, each frame of the video must be compared its the previous frame. Any differences
between the frames can be considered movement. After all instance of movement have been
detected, a list of rectangles will be created. These rectangles surround local areas of movement,
for instance a person moving in a hallway. After receiving the list of rectangles, the largest
rectangle (based upon surface area) will be displayed on the video, as seen in Figure 4. A person
measuring 70 inches in height and standing less than 50 feet from the camera will always have a
surface area greater than 20 pixels2. Therefore, the video will never display movement measuring
less than 20 pixels2. After displaying the rectangle, the system will mark the object as a
significant source of movement.

Safe Parameters
After marking a significant source of movement, the system checks for three safe
parameters: a period of no significant sources of movement, the correct infrared signal, or an
instance of the safe color located within the significant source of movement. If any of the three
safe parameters are satisfied, the servo motors will go to their safe position, equivalent to a

12

turret pointing towards the ceiling. This effectively keeps allied personnel from receiving
unnecessary harm in the case of an accidental misfire.
As each frame is analyzed for significant sources of movement, the system will monitor
how many frames do not detect movement. If motion is not detected in five successive frames,
the system assumes any targets have either left the area of interest or have been incapacitated.
The turret will then move to the safe position until movement is detected again.

IFF Color
Whenever a rectangle is displayed on the video, a small square is also placed at the
(height/3, width/2) position, as seen in figure 3. This marks the target area for the turret, as well
as the area to check for the safe color. Once the target is marked, the RGB values of the center
pixel are compared to the safe color. For example, when the safe color is blue, blue is required to
be the dominant color in the pixel. If the blue value (on a range of 0:255) is greater than both the
red and green values, then the target is deemed an authorized personnel. If the RGB values meet
the specified safe range, then the servos are set to the safe position until the safe color is no
longer detected.

IFF IR Receiver
The infrared recognition system consists of three vital components: the Sony blu-ray
remote, the IR receiver connected to the Arduino Duemilanove board, and the SimpleSerial C#
class in the motion solution. The Sony remote sends an IR signal to the IR receiver. As the
signal is received, the Arduino board records the signals and outputs through the serial port to the
SimpleSerial class.

13

All Sony remotes have the same infrared signal protocol. The signal is 12 pulses long and
follows a 2.4ms LOW pulse, as seen in Figure 6 [4]. For the signal, a pulse is considered to be a
LOW point in the signal. The code is sent in binary, so a 1.2 ms pulse is a 1 and a 0.6 ms pulse
is a 0. Each data pulse is separated by a 0.6 ms HIGH signal. The entire signal lasts for 45 ms,
and continuously repeating signals are separated by 40 ms.

Figure 6. Sony Remote Protocol

Figure 7. Arduino Board Configuration

As seen in Figure 7, a Panasonic PNA4601M infrared receiver is connected to the
Arduino board through the 5V, ground, and Digital 2 pins. A 220 resistor is connected to the
data signal in order to prevent voltage surges from passing through. A LED is connected to the
Digital ground and 13 pins, and a USB connects the board to the laptop, which also delivers
power to the Arduino board.
The Arduino Duemilanove board is coded to receive a binary signal from the Sony
remote and send the decimal equivalent number through the serial port to the computer (code can

14

be found on the website [3]). The Arduino waits for any IR signal to be sent through the IR
receiver. It measures 13 pulses (the initial 2.4 ms pulse, and the subsequent 12 pulses) and
records the length of each pulse in an array. If 13 pulses are not received (the signal is
interrupted), the code will not send partial numbers. Instead, it will send a 0 through the serial
port.
Once the signal is received, the LED will flash, letting the user know a signal has been
received from the remote. Every pulse is converted into binary: a pulse lasting longer than 1 ms
is a 1, a pulse lasting longer than 0.4 ms and less than 1 ms is a 0, and all others are a 2,
signifying invalid signals. The last 12 digits represent an individual button on the remote. For
example, a 1010011101 represents the large center button on the Sony blu-ray remote, as seen
in Figure 8. The Arduino board then converts the binary number into decimal (1010011101
becomes 1341). It prints this number on a new line and sends it through the serial port. As long
as the board is receiving power, it will always search for signals and send them through the serial
port.

Figure 8. Sony Blu-ray Remote

Buttons
In order to connect to the Arduino Duemilanove board, the laptop must have the Future
Technology Devices International (FTDI) drivers installed. The SimpleSerial class reads data
through a user selected serial port, in this case COM6. It compares incoming number to a
preselected safe number. If the user is pushing the correct button, SimpleSerial will return

15

true. If the user is pushing the incorrect button or is not pushing a button, the code will return
false.
SimpleSerial uses ReadLine( ) to receive the entire decimal number from the Arduino. If
the Arduino board is not sending a number through the serial port, the code will timeout after 10
ms and return false. Next the code compares the incoming number to a preselected safe
number, in this case 1341. If the user is pressing the correct button, SimpleSerial will return
true. If the motion code receives a true, then the servos will move to the safe position.
They will remain at the safe position until an IR signal is no longer received or the user presses
the incorrect button.

Servo Control and Calculations

If none of the safe parameters are met, the program will command the servos to engage
the target (using a laser point to measure the accuracy of the motors). In order to connect to the
1061 Advanced Servo 8-Motor control board, the Phidget Framework must be installed on the
laptop [5]. The system creates a new AdvancedServo class, which is used to control the two
servo motors. The servo motors are attached to pins 0 and 2of the 1061 control board, which are
used for altitude and azimuth control, respectively.
In order to successfully engage the target, the system must calculate the necessary angles
to send to the servo motors. The small red square shown in Figure 3 is treated as the aim point.
The horizontal and vertical distances are measured as the difference between the center of the
frame and the aim point. The depth of the object is estimated from the height of the movement
being tracked. For example, if the object is 42 pixels in height, it is estimated to be 40 feet away.
Once the depth is determined, the horizontal and vertical distances are converted from pixels to

16

feet using a conversion ratio. For example, an object 40 feet away has a ratio of 9.5 pixels to 1
foot. To determine the angles for the servos, the tan-1(x/z) and

tan-1(y/z) are calculated. In order

to move the servos, these angles are then added to the default position of the servo motors. The
default position points the servos at a 90 angle away from the box, in the same direction as the
camera; this position is equivalent to (115, 95) on the servo motors. The calculations for an
object 30 feet away can be seen in Figure 9.

Figure 9. Angle calculations

Once the angles have been determined, the servos are set to their proper position, as
shown in Figure 10. If the servos should be in the safe position, then they are set to (110, 75),
which points the servos to the ceiling. If the system attempts to set the servos to an angle that is
out of bounds, the 1061 Advanced board will catch the error and prevent the servos from
extending past their possible ranges.

Figure 10. Servo Position Function

17

4.2 Codes and Standards

The 1061 Phidget Advanced Servo 8-Motor control board utilizes a USB interface. The
Universal Serial Bus, or USB, is available in almost every computer. The USB utilizes 5V of
power, the same as the control board [6]. The webcam also uses a USB interface. The code will
be written using the C programming language. The language can be used to connect to almost
any port on a laptop.

5. Schedule, Tasks, and Milestones

While some project tasks were completed as a team (research, testing) the majority of
developmental work will be divided into two fronts; the motion tracking software and the
physical turret apparatus. The software tasks will be headed by Kevin Neas and Joleon Pettway,
while tasks related to the development of the turret will be fulfilled by Patrick Croom and
Anthony Ogidi. The WBS (Work Breakdown Statement) is found in Table 2. The entire project
schedule is displayed in Figure 11.

18

Table 2. Work Breakdown Statement

WBS Task Description

Research existing
1.1
technologies and markets

Est.
Work
Hours Owner

Resources

10

PC,KN,AO,JP Library, Internet

10
21

PC,KN,AO,JP Library, Internet

PC,KN,AO,JP Library, Internet

52

KN, JP

Visual Studio

24

KN,JP

Visual Studio

Graphical User Interface

15
Enable turret tracker code
to identify friendlies based
on color
15

KN,JP

Visual Studio

JP

Visual Studio

15

KN

Visual Studio

15

PC,AO

7.1

IR Remote code
Order parts for actual
turret
Assemble turret stand and
install servo motors and
power supply
Integrate physical turret
with Turret Tracker
Software

7.2

1.2
1.3
2.1
2.2

2.3

3
4
6.1

Technical Review Paper

Construct Design Proposal
Write motion detection
code
Write motion tracking
code

18

PC,AO

Servo Motors,laptop,
webcam, stand
materials

PC,KN,AO,JP

Visual Studio

Integrate IR and IFF code

PC,KN,AO,JP Visual Studio

7.3

Calibration

21

PC,KN,AO,JP Visual Studio

Testing

18

PC,KN,AO,JP Visual Studio

6.2

M&S

$340

19

Figure 11. Gantt Chart detailing prototype development schedule.

20

6. Acceptance Testing and Results

In order to properly simulate the systems optimum operational scenario, primary
acceptance testing was conducted in central hallway of the fourth floor of the VanLeer
classroom building. The basic setup of the tests is shown in Figure 12. There were two
phases to the final acceptance testing of the FAST system prototype. The first phase of
testing focused on the prototypes ability to successfully engage viable targets with its
field-of-fire, including the systems target differentiation capabilities in multiple target
scenarios. The second testing phase was focused on the systems integral IFF
capabilities, both the color recognition system and the IR remote transmitter. For all
acceptance tests, the system was set to operate fully autonomously; there wasnt any
human control over the prototype.

Figure 12. Basic prototype acceptance testing setup.

21

Target Engagement Tests

Single Target: Linear Path In this test a single target would approach the FAST
system with a linear path in a hallway from beyond 40 ft. The FAST systems success in
this test is determined by whether or not the system can successfully engage the target in
less than 2 seconds after the target is within 40 ft of the system.
Single Target: Erratic Path- In this test a single target would approach the FAST
system with an erratic path in a hallway from beyond 40 ft. Again, the FAST systems
success in this test is determined by whether or not the system can successfully engage
the target in less than 2 seconds after the target is within 40 ft of the system.
Multiple Targets/Target Priority In this test the FAST system had to engage two
targets in a hallway, one approaching the system from beyond 40 ft and one target
moving away from the system starting just beyond the systems field of view. The success
of the Fast system in this test is determined by whether the system detects both targets
and if the system successfully tracks only the target with the closest proximity to the
system.
Safety Function Tests
Safe Passage with IR Remote Single target will attempt to navigate a hallway
fitted with the FAST system while pressing the correct safety button on the IR remote.
Test deemed a success only if the FAST system does not engage the target, but instead
sets itself to the safe position for as long as the correct safe button is pressed on the
remote.
Safe Passage with Proper Color Single target will attempt to navigate a hallway
where the FAST system is emplaced while wearing the correct safe color. Test deemed

22

a success only if the FAST system does not engage the target and instead sets itself to the
safe position while the target is wearing the safe color.
Safe Passage with Proper Color and IR Remote Single target will attempt to
navigate a hallway fitted with the FAST system while wearing the safe and pressing the
correct button on the IR remote. The purpose of this test is to demonstrate that the safety
systems do not interfere with one another. Test deemed a success only if the FAST
system does not engage the target and sets itself to the safe position while the safe button
is pressed on the IR remote and/or the target is wearing the safe color.
Escorting Two targets will attempt to navigate a hallway equipped with the
FAST system while one is pressing the correct safe button on the IR remote and the
other is without any safety. This test is deemed a success if the FAST system does not
engage either target and sets itself to the safe position for so long as the safe button is
pressed on the IR remote.
Results
The FAST system successfully passed all of the target engagement and safety
function tests. In the single target tests the FAST systems successfully detected and
targeted the single subject whether the subject was moving in a linear or erratic path. The
FAST system on average was able to detect and track targets within 1.0 seconds of the
target entering the FAST systems 40ft. range. When faced with multiple targets the
FAST system successfully prioritized targets, engaging the target with the closest
proximity first. In the safety functions tests, the FAST system did not engage friendly
targets after detecting a correct safety parameter (correct color or IR signal) and moved
into safe position. The FAST system also successfully demonstrated the ability to

23

allow multiple targets to be escorted while a safe signal is being emitted from an IR
remote.

7. Marketing and Cost Analysis

Building a cost effective sentry gun will attract many clients including the United
States military, private security contractors, and high-tech secure facilities. Current sentry
guns are built with big and heavy equipment. The Phalanx CIWS is one of the earliest
built sentry guns. It weighs approximately 14,500lbs and cost approximately $5,000,000
USD for a single unit which is why sentry guns are currently not mass produced [7].
Other existing sentry guns have the same cost and weight problem. The FAST system is
unique by being portable, having a fast set up time, and targeting only enemies while
avoiding negligent discharge at allied forces. This prototype will significantly cut the cost
of building a sentry gun which will make it viable for mass production. The sentry gun is
a newly developed system with no similar products on the market performing the same
functions at a similar cost.
Table 3 shows the parts required to build the sentry gun prototype, as well as the
cost of each item. The prototype cost refers to how much it cost to build the actual
prototype, considering the fact that most of the items required were already available.
The actual cost is how much it would cost to build the entire unit, buying every
individual item needed. The personnel cost is based on the assumption that four engineers
worked on the project at an estimated 40 hours a week, each getting paid $30 an hour, for
a duration of 12 weeks.

24

Table 3. Estimated Design and Development Costs.

Item
Quantity Prototype Cost
(Cost to design team)
Servo Motors
2
$50
Mount
1
$50
Mini lap top
1
$0
Webcam
1
$0
Software
1
$0
Laser Marker
1
$20
Arduino
1
$0
12V Battery
1
$0
USB (x2)
1
$0
Wires
1
$0
IR
1
$0
Box
1
$0
Remote
1
$0
Personnel Cost
4
$0
Overhead (100%)
_
$170
Totals
_
$340

Estimated
Actual Cost
$45
$45
$500
$50
$200
$20
$30
$140
$7
$3
$1.50
$25
$20
$58000
$720
$59806.50

As can be seen most of the items needed to build the sentry gun are items owned by the
average American and can be bought at any local electronic store.

8. Conclusions and Final Work

Currently, the prototype for the FAST system is completed. However, several
improvements could be implemented in order to make the system better. For example:
Scaling up the design to support the weight of a machine gun
Adding a higher resolution camera
Faster processor
Facial recognition IFF capability
Incorporating RFID IFF capability
360 field view

25

 Adjusting the prototype for outdoor use

Several things could be changed to make the system work more accurately and
efficiently. Using two cameras instead of one can greatly increase the precision of motion
detection as well as exponentially increase the current range of 40ft. The size of the
housing could also be reduced to a more portable size considering the fact that only 30%
of the internal space of the current housing is used.
The initial specifications suggested a 50ft effective range, a 45 field view, and a
maximum target acquisition time of 2 seconds. We were forced to reduce the effective
range to 40ft due to the implementation of the IR receiver which had a maximum range
of 45ft. The field view was increased to 65 to compensate for the width of the hallway
which required the wider field view in order to aim for objects at a closer range. The
target acquisition rate was less than a second because the code ran more efficiently than
anticipated and implementing color recognition did not significantly slow down the code.
The initial specifications did not include the use of an IR receiver as a safety system, but
was implemented to compensate for friends not wearing a safety uniform.

26

9. References
[1] AForge.Net Framework. AForge.Net, 2010. [Online]. Available:
http://www.aforgenet.com/framework/
[2] A. Kirillov. Motion Detection Algorithms. The Code Project, 2007. [Online].
Available: http://www.codeproject.com/KB/audio-video/Motion_Detection.aspx
[3] P. Croom, K. Neas, A. Ogidi, and J. Pettway. Turret Trackers. 2010. [Online].
Available:
http://www.ece.gatech.edu/academic/courses/ece4007/10spring/ECE4007L01/dk6/
[4] D. David. SONY infrared remote protocol. Telenet, 2007. [Online]. Available:
http://users.telenet.be/davshomepage/sony.htm
[5] Drivers. Phidgets, 2010. [Online]. Available: http://www.phidgets.com/drivers.php
[6] Universal Serial Bus Specification, Rev. 3.0, 2008. [Online]. Available:
http://www.usb.org/developers/docsyeah
[7] Phalanx Close-In Weapon System, Raytheon Company, Waltham, MA, 2008.
[Online]. Available:
http://www.raytheon.com/capabilities/rtnwcm/groups/rms/documents/content/rtn_rms
_ps_phalanx_ciws_datash.pdf

27