Object Detection and Tracking using Turtle Bot

About: During a 6-month hybrid internship at IISER Bhopal MoonLab, under the guidance of Professor P. B. Sujit, we successfully created an object tracking Rover. The internship consisted of 2 months of offline work followed by 4 months of online collaboration. The Rover is equipped with YOLO V3 for detecting objects in video stream data, and we employed the Kalman Filter to process this data. The filtered output was then used as input for a custom PID controller, which we fine-tuned based on the camera's positional error and the estimated velocity of the moving object. To validate the system, we conducted tests using Webots simulation and Turtlebot 3 Burger, and the results have been generated.

TurtleBot 3 : Burger

LoginID & Password

moonlab
iiserb@1234

ROS Execution Commands

1. Vision

roslaunch drive_rover vision.launch

2. Drive

roslaunch drive_rover drive.launch

3. Bringup

source ros/turtlebot/devel/setup.bash

export OPENCR_MODEL=burger

export OPENCR_PORT=/dev/ttyACM0

sudo chmod a+rw /dev/ttyACM0

cd ros/opencr_update/

./update.sh $OPENCR_PORT $OPENCR_MODEL.opencr

roslaunch turtlebot3_bringup turtlebot3_robot.launch 

4. Teleop

source ros/turtlebot/devel/setup.bash

roslaunch turtlebot3_teleop turtlebot3_teleop_key.launch

5. SLAM

source ros/turtlebot/devel/setup.bash

export TURTLEBOT3_MODEL=burger

export LDS_MODEL=LDS-01

roslaunch turtlebot3_slam turtlebot3_slam.launch

Results

1. Simulation

Simulation.mov

2. Hardware Implementation

IISER-B.Rover.mp4

Turtlebot_Bottle_Tracking.mp4

Collaborators:

Abhimanyu Bhowmik, Madhushree Sannigrahi, Dr. P.B. Sujit

Links:

MoonLab@IISER-B