GUJARAT COUNCIL ON SCIENCE AND TECHNOLOGY
Dept. of Science and Technology, Govt. of Gujarat
Application format for the
IDEATION STAGE FOR ROBOT MAKING COMPETITION
ROBOFEST GUJARAT-5.0
Application No.:
1. Proposal submitted for Robot: (Please tick mark the category of robot proposed)
• For Senior Level:
1. Autonomous Maze solver
2. Type of Institution:
Government: ☐
3. Name of School/Institute/Department/Organization:
Faculty of Technology and Engineering MSU Baroda
4. Complete Postal Address with Pin code:
Kalabhavan, Palace RD, Nr. Kirtistambh, Babajipura, Vadodara, Gujarat 390001
5. Name of affiliated University/Board Name:
The Maharaja Sayajirao University of Baroda
6. Mentor/Coach/Faculty Guide Name: Jyoti Sohoni
7. Designation of Mentor/Coach/Faculty Guide: Assistant Professor (Temporary)
8. Experience of Mentor/Coach/Faculty in years: 17
9. Email address of mentor: [Link]-eed@[Link]
10. Mobile Contact No of mentor: +919979953605
11. Office No: +919979953605
1
12. Proposed Team of Making robot: (Maximum four participant per team allowed)
Name of E-mail address
Sr. Institute Programme Standard/
Student/Research Department Mobile No.
No. Name enrolled Year
scholar
1 Nisarg Brijen Faculty of Electrical Electrical 3rd 7016323520 goodnisarg@[Link]
Thakar Technology Engineering
and
Engineering,
MSU
2 Shivam Mohan Faculty of Electrical Electrical 3rd 8128534925 mohan2005singh12@[Link]
Singh Technology Engineering
and
Engineering,
MSU
3 Devshree Hitesh Faculty of Electronics Electronics 3rd 9712751955 devshreehalbe18@[Link]
Halbe Technology Engineering
and
Engineering,
MSU
4 Yug Faculty of Mechanical Mechanical 3rd 9327791414 Kaharyug17@[Link]
Khumeshar bhai Technology Engineering
Kahar and
Engineering,
MSU
13. Name of the authority in whose name Cheque / Demand Draft should be drawn. (Institutional Bank
account only. Do not enter mentor or student bank details)
Name of the Account Holder: Dean Faculty of Technology & Engineering
Name of the Bank: Bank of Baroda, Dandia Bazar Branch
Bank Account No: 05780200000021
IFSC Code: BARD0DANDIA
MICR Code: 390012004
2
3
4
IDEATION DOCUMENT FOR
MAZE SOLVING BOT
Faculty of Technology and Engineering
MSU Baroda
Robotics & AV Club
1. Type of Robot: Autonomous Maze Solver
2. Robot Assembly Design:
FRONT VIEW
5
TOP VIEW
SIDE VIEW
6
Electrical Assembly
• Motors are required to drive the bot, N20 series motor with encoder are used which
would enable it to give the data of length of path traversed to Microcontroller.
• LiDAR for measuring the distances in all four directions and to be able to make a
map of maze using point clouds.
• IMU for knowing the orientation of the bot when it taking turns so the map and
algorithm can be updated accordingly.
• Raspberry Pi Pico 2 microcontroller is used which would take decisions &
integrate the data from LiDAR to make a map of the maze and save it in non-
volatile memory.
7
• A motor driver is required to control the motors since they operate at a higher
voltage compared to Microcontroller. It would take signals from Microcontroller
and operate the motor accordingly.
• A 12V rechargeable battery is used as a power source and thus a buck is required
for stepping down the voltage to 5V for supplying power to Microcontroller.
• A PCB is used rather than connecting wires physically to make design robust and
clean.
8
1)TOP VIEW OF PCB
9
2)BACK VIEW OF PCB
10
Mechanical Assembly
Two motor and one castor assembly is used. The motors selected are geared motors
which have high torque and low rpm.
A frame is required on which the components can be mounted according to the
measurements to make precise and accurate measurements position and location of bot
and to detect a correct error for the control system.
This is important since the calculations made on the data from the LiDAR to take
decisions and make map must be reliable and exact.
House-hold materials can be used ensuring they meet the stress-strain & other design
criteria in addition to some specially manufactured materials to make the body of the
bot.
Control System Consideration
LQR (Linear Quadratic Regulator) for a maze solver bot is used which is a more
advanced approach than using simple PID. While PID controls only one variable at a
time, LQR can handle multiple states simultaneously—like position, angle, and
velocity—while minimizing both error and control effort in an optimal way.
Source: ResearchGate
11
3. List of all components
Types of components Name of Components
Structural Components • Chassis
Motional Components • Wheels (Diameter 34mm) and
Castor wheel
• N20 motors
Electronics Components • Raspberry Pi Pico 2
• TB6612FNG
• LiDAR sensor (Vl53l0x)
• XL6009 Buck Converter
• Lithium Polymer Battery
• MPU9250 9-Axis Attitude Gyro
Accelerator Magnetometer Sensor
Module
4. The Methodology of Making Robot
Defining the Problem:
A maze is given. It is as an arrangement of blocks such that it makes a network of
paths that has (for our case) multiple entry and exit points.
Source: Python Point Cloud Source: Azim Premji University
12
We strive to make a bot using an interdisciplinary approach of Electrical, Mechanical
and Software fields taking reference from theoretical knowledge from Books and
execute it that would find a way out of the maze from any starting point in maze and
would also make a map using LiDAR sensor’s point cloud data using Python.
The final goal would be to make a maze bot that must independently navigate more
than one complex mazes (placed in cascaded fashion) with multiple decision points,
dead-ends, and obstacles, without any external control or preloaded map in a dynamic
and unpredictable environment.
The bot is required to be fully autonomous, have autonomous decision making and
real time path planning.
Understanding the Electrical Hardware Working:
The first thing that is most important is to move the design made with a power source.
This task would be achieved by motors.
13
The second would be a sensor which would take the data of surrounding such
distance of the wall using LIDAR.
LIDAR, which stands for Light Detection and Ranging, is a remote sensing method
that uses light in the form of a pulsed laser to measure ranges (variable distances).
Source: [Link]
14
The third is a Raspberry Pi Pico 2 Microcontroller which would take decisions on data
inputs provided by the sensors based on algorithm. It also has non-volatile memory
which would store the maps of the maze made by bot and the algorithm’s code itself.
The fourth requirement is a motor driver since the Microcontroller operates at a low
voltage(5V).
It would link control signals from microcontroller at 5V to motor p at 11-12V.
15
The fifth requirement is a IMU Sensor which would provide the orientation of the bot
in the maze.
16
Looking for ways to solve a Maze Algorithmically and Intuitively
For Exploration run from a random start –BREADTH FIRST SEARCH
(BFS)
BFS is algorithm used for traversal of graphs by exploring all nodes at the current level
before moving on to node at the next depth level. The core idea of BFS is to expand outward
from a starting node, visiting its neighbours, then visiting the neighbours of those
neighbours, and so on until it reaches the final goal.
WORKING:
Initialization:
▪ Choose a starting node.
▪ Create a queue and add the starting node to it.
▪ Mark the starting node as visited to prevent revisiting it later
▪ Create a Boolean array to keep track of visited nodes.
Traversal:
▪ While the queue is not empty:
• Dequeue a node from the front of the queue. This is the current node
being processed.
• For each unvisited neighbour of the current node:
Mark the neighbour as visited.
17
Enqueue the neighbour into the queue.
Completion:
▪ The process continues until the queue becomes empty, indicating that all
reachable nodes from the starting node have been visited.
For Optimized shortest path run- A star
It is Informed Type Algorithm which finds shortest path by taking in account location of
goal while deciding next move, towards the goal using heuristics
WORKING PRINCIPLE
F(n)= G(n)+H(n)
• G(n): Actual cost from the start node to the current node n
• H(n): Heuristic estimate of the cost from n to the goal
• F(n): Total estimated cost of the cheapest solution through n
WORKING
1. Initialize the priority queue with the start node and its heuristic value.
2. Loop until the priority queue is empty or the goal node is found:
a. Pop the node with the minimum f-value (f = g + h) from the priority queue.
b. If the popped node is the goal node, stop the search.
c. Otherwise, expand the node and update the priority queue with the neighbouring
nodes and their f-values.
3. If the goal node is found, reconstruct the path.
18
Making Maps using Point Clouds & Numerical Methods
Source: Numerical Methods for Engineers [3]
Source: Wikipedia
19
Point clouds are the group of points having coordinates (x,y,z) and even a time stamp,
processed through the data given by the LiDAR.
These can be further used to make a 2d or 3d maps of the surroundings using Point Cloud
Techniques.
These techniques can be implemented using python language and the Raspberry Pi Pico 2
which is compatible with python.
2D or 3D maps of the maze can be made and saved in the non-volatile memory, which can
be processed for the next run from which an optimal path can be found using numerical
methods.
20
5. Application of Proposed Robot in Societal Context
The Unique Idea
Public Places and Religious Places where there are chances of Stampede and
Crowd
Rush and stampede can occur in the context of India in fairs, concerts, celebrations,
railway stations, bus stations and religious places.
The Maze solver Bot can be used to avoid and rescue by developing a system in
following parts.
1. Stationery Sensors for Feeding Movements and Flow of Public
2. Synchronised Moving Bots which communicate with each other and input data
from Stationery Sensors to find way out/in.
Other Uses
A. Pipeline and Sewer Inspection (Hazardous for Humans due to toxic gases).
B. Search and Rescue Operations in Earthquakes, Storms and Fire.
C. Military Operations
21
6. Size of robot proposed for Proof of concept :
a. Length __10_cm_____
b. Width __ 10_cm______
c . Height ___5_cm______
22
7. Timeline for Robot making with milestones.
23
8.A photograph serving as proposed outline for understanding of evaluation
committee.
Top View
Side View
24
Front view
BFS Algorithm Point Cloud Representation
Source: Point Cloud Library [1]
25
Bibliography:
Point Clouds
[Link] [1]
[Link] [2]
Books for reference that we think would be helpful for the project:
1. Numerical Methods for Engineers
by Raymond Canale and Steven C. Chapra [3]
2. Control Systems [4]
by M. Gopal
3. Introduction to Algorithms [5]
by Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest, and Clifford Stein
26