Neural Network-Based Compensation Control of Robot Manipulators with Unknown Dynamics

Abstract

A neural network (NN)-based compensation control is proposed for the trajectory tracking of robotic manipulators with unknown dynamics. This compensation controller includes a PD feedback controller, a nonlinear feedback controller and a neural network compensator with input modification. The PD controller and the nonlinear feedback controller are used to ensure the stability of the robot system, while the neural network is employed to provide the required feedforward compensation input torque for the trajectory tracking. The inputs of the neural network are modified by the reference tracking error and the network derivatives in order to further improve the control performance. The theoretical result concerning the tracking error asymptotically convergent to a neighborhood of zero is given. In addition, a NN-based robust compensation controller is proposed, using a slide mode control in the control law, which leads to asymptotic stability of the tracking errors. Simulation studies have been carried out to verify the effectiveness of the approaches and show the feasibility of the proposed schemes on a two-link manipulator.