Path Planning in a dynamic indoor environment for mobile robots using Q-Learning Technique

Abstract

Autonomous Navigation for mobile robots has many applications for indoor and outdoor environments; however, it is still a challenge since no error-free solution for its implementation exists yet. This study attempts to present a path planning approach using Q-learning, a Reinforcement Learning technique, to be deployed and tested in a simulated warehouse-like environment. The approach used was able to generate a collision-free path for the robot to navigate through.