A Reinforcement-Learning-Based Adaptive Sliding Mode Controller for Robotic Manipulators

Abstract

In this paper, a reinforcement-learning-based adaptive sliding mode controller (RLASMC) is proposed to achieve more precise tracking control in robotic manipulator systems with nonlinear friction, modeling errors, and external disturbances. In this controller, a robust term is designed to compensate for the external disturbance, system uncertainty, and joint friction. Furthermore, the dynamic information of the robotic manipulator is employed as input to a reinforcement learning agent, enabling the agent to optimize the parameters of the sliding mode controller within a continuous action space. Simulation studies are implemented to validate the effectiveness of the proposed controller.