<![CDATA[This article presents a robust optimal tracking control approach for a Four Mecanum Wheeled Robot (FMWR) using an online actor-critic reinforcement learning (RL) algorithm to address the challenge of precise trajectory tracking problem in the presence of mass eccentricity and friction uncertainty. In order to handle these obstacles, a detailed dynamics model is derived using Lagrange’s equation, and the Hamilton–Jacobi–Bellman (HJB) equation is solved by iteration algorithm with policy evaluation and improvement. The training laws of optimal control law and value function are proposed after minimizing the modified Hamiltonian function. Moreover, to handle the time-varying property of tracking error model, a transform is given with the addition of time derivative term. Simulation Studies demonstrate the approach’s effectiveness, significantly improving trajectory tracking accuracy and robustness against disturbances. This research contributes to mobile robotics by enhancing control precision and reliability in dynamic environments.]]>
Copyrights © 2024
