Path Planning and Control for Omni-directional Mobile Robot using Q-Learning and CLIK Algorithm for Home Environment

Abstract

Localization, mapping, and planning are the three crucial steps to accomplishing the autonomous navigation of mobile robots in an unfamiliar environment. Since implementation of reinforcement learning (RL) algorithms for autonomous navigation in the case of omni-directional robots is a less explored research area, and also such robots have a unique feature over differential drive robots that they can also produce sideways movement. Therefore, in this paper, an RL algorithm called Q-learning is used to get the safe and shortest path from a start point (SP) to a goal point (GP) in a home environment. The path trajectories are obtained by using polynomial curve fitting. The closed-loop inverse kinematics (CLIK) algorithm is used to control a three-wheeled omni-directional mobile robot to follow the desired path. The simulation and plotting are done using MATLAB. The simulation results show that the suggested algorithm can effectively recognize and avoid static obstacles of different shapes and dimensions in an indoor home environment.