Treffer: Automated Car Guiding System Using Reinforcement Learning

The major objective of this project is to design and implement a car guiding system in a desk-size area, with a remote-controlled toy car. The software, including the calculation, image processing, and movement control, was coded with Python and C++. Q-learning algorithm was selected to be the core of the calculation part, and OpenCV library was used for image processing.The hardware consists of a webcam, a laptop, and a toy car with Bluetooth connection. Some wood boards were also used to build a frame for restraining the area for running the car.The system is able to track the car, detect the obstacles, calculate the optimal path, and send signals to the car in order to control the movement.