PERIMETER SURVEILLANCE FOR SENSITIVE

LOCATIONS USING AUTONOMOUS SENTRY

ROBOT

A project report submitted for the partial fulfilment of academic

requirements for the award of Degree in Bachelor of Engineering in the
Department of Electrical and Electronics Engineering

Submitted by

GURURAJA K S 4NI19EE030
M R SURYA RAJU 4NI19EE403
NITHIN S 4NI19EE069
SALMAN J K 4NI19EE088

Under the supervision of

Miss POOJA A R
Assistant Professor

Department of Electrical and Electronics Engineering

The National Institute of Engineering
(An Autonomous Institute under Visvesvaraya Technological University,
Belagavi)

Manandavadi Road, Mysuru – 570 008, Karnataka, INDIA

January 2023
 ABSTRACT

“With the advent of Asymmetric Warfare, rapidly deployable hardware is key for
success in the modern battlefield”. This is the hard fact that governs most of all
modern battles being fought around the world today.

Going by this design philosophy, an attempt is being made to design and construct
an Autonomous Sentry Robot that provides surveillance capability for sensitive
installations, such as critical Industries and Defence establishments.

This Robot will be modular and can easily be deployed on existing physical
infrastructure, the system aims to provide maximum security features whilst
adhering to the concept of Asymmetric Modern Warfare.
 INTRODUCTION

Leveraging inferior tactical or operational strength against the vulnerabilities of a

superior opponent to achieve disproportionate effect with the aim of undermining
the opponent's will in order to achieve the mission requirement is the asymmetric
actor’s strategic objective.

This is being proven in conflict zones such as Syria, Iraq and Israel, closer to home,
the effects can also be seen at POK. There was a time when countries with high
military budgets would have tried to either overwhelm or outgun the enemy forces,
but even countries like the United States of America are shifting to a more
calculated approach to address modern threats in the battle field. The prototype
sentry robot being presented is a step in the same direction.
 LITERATURE SURVEY

1. Design and Implementation of Brushless The paper designs a high power brushless DC
DC Motor Drive and Control System motor (BLDC) closed-loop control system,
Huazhang WAN. mainly including design of H bridge drive
circuit, Control of rotation direction for motor
and speed detection circuit in order to
improve the performance of motor running,
one adopts PID algorithm, through the tuning
of parameters, the control exhibits very good
performance. Experiments show that both
hardware and software control algorithms are
reliable, stable. The running performance of
the system is robust before or after adding
load.

2. Modelling and Simulation of the BLDC This paper presents the development of
Motor in MATLAB GUI Balogh Tibor,
Viliam Fedák, František Ďurovský. control and power electronics parts for a
BLDC motor drive. The signals from the
most commonly used sensors for this
application are used. The first one is the
incremental encoder for speed and relative
position detection. The contemporary three
phase bridge driver integrated circuit used in
this application provides good protection
functionality to protect the bridge and the
motor from excessive currents and short-
circuit conditions.

3. BLDC Motor Driver – Development of The paper proposes a model of brushless DC

Control and Power Electronics D. Krklješ, motor considering behaviour of the motor
Morvai, K.Babkoviü and L.Nagy. during commutation. The torque
characteristic of BLDC motor presents a very
important factor in design of the BLDC
motor drive system, so it is necessary to
predict the precise value of torque, which is
determined by the waveforms of back-EMF.

4. I m a g e b a s e d P a n - T i l t c o n t r o l i n Position of an object detected in the

a multi-camera surveillance
e n v i r o n m e n t Ser-Nam Lim, Ahmed zero-positions while different pan-tilt
Elgammal and Larry S. Davis positions of the same camera are
related in the form of projective
rotations. We then derive that the
trajectories in the image plane
corresponding to these projective
 rotations are approximately circular
for pan and linear for tilt. The camera
control technique is subsequently
tested in a working prototype one
camera is related to the other cameras by
homographies between the zero-positions
while different pan-tilt positions of the same
camera are related in the form of projective
rotations.

5. Face detection and recognition using This paper explores the feasibility of
Raspberry-Pi.Ishita Gupta, Mrs. Varsha implementing Raspberry Pi based face
Patil, Chaitali Kadam, Shreya Dumbre recognition system using conventional face
detection and recognition techniques such as
HAAR detection and PCA.

6. Object Detection and Human Identification This paper is to investigates the development
using Raspberry Pi. Dr Sunitha M R of a navigation system for blind and visually
Fathima Khan impaired people along with human
identification. It is based on a Raspberry-Pi
with synthetic speech output. This paper
works onto improve HAAR – Like
classification’s accuracy and also get the
most in real – time performance, by using an
algorithm called AdaBoost – Adaptive Boost
Learning Algorithm, that can adaptively
select each feature and boost it in each step
and them combine them to form a stronger
classifier.

7. Face Detection and Identification using YOLO (You Only Look Once) is a state of
YOLO M S Antony Vigil Manasi the art object detection system. The object
Makarand Barhanpurkar NS Rahul detection algorithms of YOLO is taken and is
Anand used on a custom made dataset to identify the
person on camera. This paper addresses
fundamental challenges faced in making the
dataset. It also compares conventional
methods with YOLO.

8. Design of Battery Management System

Rashmi Ravi Kumar and Ratan Soni
 PROBLEM STATEMENT
• Due to the sheer span of the LOC (Line of Control) manual surveillance is
extremely challenging.
• During monsoons, many sections of the LOC get flooded forcing soldiers to
wade in knee deep water.
• After a search party passes by a section, the section is left unattended until
the next rounds.
• Having static CCTV cameras on fence posts at regular intervals is not
sufficient as it won't provide enough intelligence on enemy moments due to
limitations in camera angles.
 OBJECTIVE

Design and Fabricate a research backed, modular Autonomous Sentry robot, which
includes object detection capabilities to assist in surveillance duties.

SCOPE

• Design and fabricate a monorail or multi-rail support structure to support

the mass of the robot as it glides along.
• Divide the robot into it’s functional subsystems and make each section
modular based on the degree of modularity required.
• Depending on the capabilities and availability of Single Board Computers
and necessary hardware peripherals, arrive to a conclusion on wether to
execute edge computing or to stream data back to a central computer at the
command centre.
• Leverage 3D printing technology to create a robotics platform which is easy
to customise in the field.
• Design and build, either a hardware or software based control application to
take charge of the robot when it is not being used in Autonomous Mode.
 METHODOLOGY

After the deliberations till this point, the following subsystems have been decided
1. Optics Module
2. SBC Module, will be substituted with control module if Edge computing is
not feasible
3. Battery Module
4. Motor Module
5. Accessories
6. Rails and physical Infrastructure.

1.Optics Module

Pan tilt mechanism is basically an idea for grasping the images or videos of
adversities or obstacle present in the hemispheric atmosphere in front to the camera
module.
The sideways or pan motion helps to view the object that is moving right to left or
vice versa, in other hand up and down or tilt motion of camera helps to view the
object that is moving up and down.
For this purpose, Servo motor is being used, which works on the principal of closed
loop system.
The code for this will be uploaded to development board, and will be made cross-
verification of expected result.
The following steps would be taken in future to alter the desired results:
 • Decision will be made whether to use stepper motor or servomotor to
operate.
• Now any development board like Arduino(since easily accessible) is used
here to get analog input from joystick and maps it to the input required for
the servos.
• In future better performing SBC which is sufficient for need will be used.

2.SBC Module

Due to global chip shortages the number of SBCs available around the world are
very limited, if available, the cost is very prohibitive. Due to the before mentioned
reasons the following choice was made.
A Raspberry Pi 4 SBC and a HD webcam (that is already in the team’s possession)
will be used to experiment with some of the object detection algorithms and gauge
the feasibility of a robot with edge computing capabilities.

The following steps will be taken in the future to decide on a control system in the
future
• Research and experiment on some of the common object detection
algorithms.
• Research and experiment on some of the popular camera choice used in
hobby grade projects and research projects.
• Analyse the situation of the global chip and SBC market and decide on a
cost effective choice for a SBC and Camera combination.
• Run calculations to determine the power usage and cooling methodologies
required to run the SBC module setup. Go through through the previous
three steps until a good match is found.
• If a suitable combination is not found the team will have to shift to using an
IP camera as a source of video feed and offload the heavier computing to a
computer that is located in the command centre. This still retains the
autonomous capability of the robot, just the computing gets offloaded via
the wireless backbone network.

After a suitable SBC system has been chosen, a shock resistant mount needs to be
designed to safely mount the SBC within the case. Appropriate shock resistant
designs needs to be made to protect the SBC from mechanical stress.
The next decision is the quick connect PCBs, these are common PCBs which will
be present in each subsystem facilitating the quick interchanges
3.BATTERY MODULE
 4.MOTOR MODULE

The job of the motor module is to make sure the motor is driven across the rail as
per the need. Depending on the cross-section i.e. the profile being used, the type of
wheels and their arrangement will be chosen.
For example a choice can be made to use a rectangular profile or a round profile for
the rails. A choice can be made to suspend the robot from the rail to keep the CG
lower or choose to have the robot ride on top of the rails. More research is needed
to take an informed decision on the matter as factors like precipitation, wind, snow
and dust need to considered before a decision is made on the rail system and decide
a drive system for it.
For sizing the motor and motor drivers, the first consideration will be the weight of
the robot and then the dimensions and CG, using these parameters a choice can be
made on the number of wheels (both driven and non driven). Taking into account
the diameter and the contact patch of the motor a calculation can be done to choose
the motor of appropriate RPM and Torque.
To determine the end of the rails, a limit detector needs to be developed. The limit
detector can range from a mechanic version like limit switches to optical limit
detectors, like infrared distance sensors. Similar to choice for wheels, a choice for
choosing the limit detector can only be taken after considering the environmental
factors in which the robot will be operating in. This is consideration is quite
important as the limit detector will be subjected to harsh working conditions.
 5.Accessories

• Depending on the type for cameras used the robot might need illuminators,
both visible light and Infra-red.
• The Robot might also need Hooters and other visual alarm aids, these aids
can also be included in the accessories module.

EXPECTED OUTCOME

• Successfully design and fabricate a modular robot that achieves the

surveillance capabilities required.
• Design modular functional units, these units shall be easily swappable on
the robot whilst also being easy to manufacture.
• Leverage 3D printing technology for quick iteration times and rapid
prototyping for modifications.
• Code in autonomy to the robot in terms of object detection and means to
report detected objects back to the control centre.
• Design in means of manual takeover / manual control when the need arises,
so as to introduce a level of Human in Loop principle when it comes to the
detection of objects.
• Design and fabricate a shorter version of the rails needed to demonstrate the
functionality of the robot.
• Design and fabricate the charging dock station needed for the robot to
charge itself at regular intervals.
 REFERENCES

1. Design and Implementation of Brushless DC Motor Drive and Control

System Huazhang WAN.
2. Modelling and Simulation of the BLDC Motor in MATLAB GUI Balogh
Tibor, Viliam Fedák, František Ďurovský.
3. BLDC Motor Driver – Development of Control and Power Electronics D.
Krklješ, Morvai, K.Babkoviü and L.Nagy.
4. I m a g e based Pan-Tilt control in a multi-camera
s u r v e i l l a n c e e n v i r o n m e n t Ser-Nam Lim, Ahmed Elgammal
and Larry S. Davis
5. Face detection and recognition using Raspberry-Pi. Ishita Gupta, Mrs.
Varsha Patil, Chaitali Kadam, Shreya Dumbre
6. Object Detection and Human Identification using Raspberry Pi. Dr Sunitha
M R Fathima Khan
7. Face Detection and Identification using YOLO M S Antony Vigil Manasi
Makarand Barhanpurkar NS Rahul Anand
8. Design of Battery Management System Rashmi Ravi Kumar and Ratan
Soni