Goalkeeper Robot using Reinforcement Learning and Computer Vision

Abstract

In this project, we aim to build an intelligent goalkeeper robot that is able to prevent the ball from passing the goal border. The robot will move in a linear direction. It will be trained by trial and error through reinforcement learning (using the Q-learning algorithm) and computer vision techniques. After training, the robot should be able to detect where the ball is and move toward the ball to prevent it from passing the goal border. The hardware components include an Arduino UNO microcontroller, linear actuator (slider with a stepper motor), motor drivers, HP laptop for processing, and a camera as an input device. Since Python is an efficient language for dealing with complex algorithms in machine learning (ML) and computer vision (CV), we are going to use it to develop the software part of the system.