Neural network based adaptive dynamic surface control for flexible-joint robots

Abstract

A neural network based adaptive dynamic surface controller is proposed for uncertain flexible-joint robot systems. The dynamic surface control method eliminates the problem of “explosion of complexity” existing in traditional backstepping approach by the addition of low pass filters. RBF neural networks are used to approximate the unknown nonlinearities of the model. Nonlinear damping items are used to overcome the external disturbances. Adaptive laws are designed to estimate the weight values of the neural networks and unknown parameters. From Lyapunov stability analysis, it is shown that the control strategy can guarantee the semi-global stability of the closed-loop system and arbitrarily small tracking error by adjusting the controller parameters. Simulation results are presented to validate the good tracking performance of the control system.