Adaptive Dynamic Programming-based Decentralized Sliding Mode Optimal Control for Modular and Reconfigurable Robots

Abstract

A model-free decentralized sliding mode control (SMC) is proposed via adaptive dynamic programming (ADP) algorithm to solve the problem of optimal tracking control of modular and reconfigurable robots (MRRs) in this paper. The dynamic formulation of MRR is expressed by a synthesis of joint subsystems with interconnected dynamic couplings (IDCs). Based on SMC technique, the optimal control of robotic system is transformed into an optimal compensation problem of unknown dynamics of each subsystem and a neural network (NN) identifier is set up to approximate IDC dynamics. Based on ADP and policy iteration (PI) method, the Hamilton-Jacobi-Bellman (HJB) equation can be addressed by using the critic NN and the optimal control policy can be obtained. The closed-loop robotic system is proved to be asymptotic stable by using the Lyapunov theory. Finally, simulation results are provided to demonstrate the effectiveness of the method.