Backstepping Control of Robot Manipulators Driven by Induction Motors Using Neural Networks

Abstract

A robust control for robot manipulators actuated by induction motors using neural networks is considered. The control is designed to compensate for nonlinear dynamics associated with the mechanical subsystem and the electrical subsystems only with the measurements of link position, link velocity and stator winding currents. Two-layer NNs are used to approximate unknown functions occurring from parameter variation during backstepping controller is designed to achieve uniform, ultimately bounded link position tracking of the given reference signal. Simulation results are presented to illustrate the tracking performance of the proposed controller applied to a one-link robot manipulator.