Decentralized adaptive tracking control of robot manipulators

Abstract

This paper presents a new decentralized adaptive scheme for trajectory tracking control of robot manipulators. The proposed controller is extremely simple computationally, does not require knowledge of the manipulator dynamic model, and is implementable in either joint-space or taskspace with nonredundant and redundant manipulators and with robots containing closed kinematic chains. It is shown that the control strategy is globally stable in the presence of bounded disturbances, and that the ultimate size of the tracking errors can be made arbitrarily small. Results of laboratory experiments with a PUMA 560 manipulator demonstrate that the proposed controller provides a simple and effective means of obtaining high performance trajectory tracking.