On passivity and simple adaptive control of nonlinear manipulators

Abstract

Extension of passivity to a class of nonlinear systems linear in control and its application to robust adaptive control of robots are considered. Passivity, also called positive realness in linear time-invariant systems, is needed for realizability of networks, optimality of control systems, and stability of nonlinear or adaptive controllers. It is shown that stabilizability and parallel feedforward can be used to satisfy passivity conditions in manipulators. This result guarantees robust stability of nonstationary, nonlinear, or adaptive controllers if they are applied to almost strictly passive (ASP) configurations.<>