Robust neural network-based tracking control for electrically driven constrained robots with constraint uncertainties

This paper addresses the problem of designing robust tracking controls for constrained robot systems actuated by brushed direct current motors. Both mechanical dynamics and electrical dynamics in the electrically driven constrained mechanical system are unknown and neural network approximation systems are constructed to learn the behaviors of these two uncertain terms. Moreover, the constraint surface can be allowed to be perturbed by time-varying bounded uncertainties. By using the backstepping technique, an adaptive neural network-based dynamic feedback tracking controller is developed such that all the states and signals of the closed-loop system are bounded and the trajectory tracking error can be made as small as possible.