Autonomous Driving Mobile Robot using Q-learning

Abstract

This study presents a Reinforcement Learning, more specifically Q-Learning for the obstacle avoidance autonomous mobile robot. For the past few years, Reinforcement Learning in Robotics has been a difficult problem to tackle. Numerous in-depth research efforts have been prompted by the opportunity to provide a robot with a tool capable of enabling the development of an ideal behavior via multiple hit and trial interactions with the robotic environment. In this paper, Implementation of Q-learning algorithm and feedback control for the mobile robot has been implemented in ROS. This algorithm is developed and tested with the turtlebot3_burger. The control script and the simulated robot communicate via ROS.