Adaptive NN impedance control of constrained mechanical systems

Abstract

Adaptive neural network (NN) controller design is presented in this paper for impedance control of uncertain mechanical systems subjected to a set of holonomic constraints. Some properties of the dynamic model are exploited to facilitate the controller design. An adaptive neural network controller is constructed in order to eliminate the need for the tedious dynamic modeling and the error prone process in obtaining the regressor matrix. The proposed controller guarantees the motion tracking control and force errors asymptotically converge to the desired manifold, and regulates the motion/force relationship to the desired impedance dynamics. Numerical simulation has been done to show the effectiveness of the proposed controller for the constrained mechanical systems.