During the Winter 2021 Introduction to Robotics class, my teammate and I implemented a Q-Learning algorithm in Python3/ROS and trained a Turtlebot3 in Gazebo to match dumbbells to numbered blocks.
This project involved a mix of robot kinematics, vision recognition, and reinforcement learning in order to train the Turtlebot.
Check out one of our test runs below:
To get more details on this project, peep the README on our corresponding GitHub repo!