Application of Deep Reinforcement Learning for Tracking Control of 3WD Omnidirectional Mobile Robot

Abstract

Deep reinforcement learning, the fastest growing technique, to solve real-world complex problems by creating
a simple mathematical framework. It includes an agent, action, environment, and a reward. An agent will interact
with the environment, takes an optimal action aiming to maximize the total reward. This paper proposes
the compelling technique of deep deterministic policy gradient for solving the complex continuous action
space of 3-wheeled omnidirectional mobile robots. Three-wheeled Omnidirectional mobile robots tracking is
a difficult task because of the orientation of the wheels which makes it rotate around its own axis rather to
follow the trajectory. A deep deterministic policy gradient (DDPG) algorithm has been designed to train in environments
with continuous action space to follow the trajectory by training the neural networks defined for
the policy and value function to maximize the reward function defined for the tracking of the trajectory. DDPG
agent environment is created in the Reinforcement learning toolbox in MATLAB 2019 while for Actor and critic
network design deep neural network designer is used. Results are shown to illustrate the effectiveness of the
technique with a convergence of error approximately to zero.