A Flexible Sliding Mode Controller for Robot Manipulators Using a New Type of Neural-Network Predictor

摘要

This paper presents a free-model high-precision controller for robot manipulators using variation sliding mode scheme and a semi-positive neural-network design. The total dynamics of the robot is first estimated by the neural network in which a new type of learning laws is proposed based on twisting excitation signals. A flexible sliding mode control interface is then developed to realize control objectives using the estimation result of the neural network. The control performance of the closed-loop system is verified by the intensively theoretical proof and simulation results.