Multivariable Root Loci of Control Systems of Robot Manipulators : 2nd Report, The Case of Dynamic Compensators

Abstract

Currently, robot manipulators are controlled by decentralized or independent joint controllers. That is, each joint is locally controlled using only local information. We discuss the asymptotic behavior of the closed loop eigenvalues of such systems in a case where the dynamic compensators are used as feedback controllers. The high-gain feedback is an effective control method because of its robustness to parameter variations and disturbances. By using theory of multivariable root loci, we present four theories about the behavior of the closed-loop eigenvalues when the high-gain feedback is implemented. It is supposed that the obtained results could be useful in the control system design of manipulators.