Decentralized robust control of robot manipulators

Abstract

Presents a robust decentralized control for robots. It follows the design procedure for a class of interconnected nonlinear mechanical systems presented in Tang and Tomizuka (1997). The rationale under this design procedure is to compensate high-order interconnections in states among subsystems. The resulting controller is extremely simple, consisting of a linear state-feedback plus a signal designed to compensate for the coupling among the joints, parameter uncertainty and bounded disturbances. The implementation of the controller needs no robot model knowledge. This feature plus its suitability for parallel processing implementation makes the proposed controller promising for a robot to perform fast-tracking tasks. Experimental results are included to illustrate the performance of this controller.