Adaptive unknown input observer for nonlinear systems with application to fault diagnosis

Abstract

An adaptive unknown input observer for nonlinear systems whose nonlinear functions satisfy Lipschitz condition is presented. The main advantage of the proposed observer is that it is sufficient to satisfy the standard existence conditions of normal unknown input observer design. Another advantage is that it estimates both states and parameters simultaneously. So it can be very much useful for robust control and fault diagnosis. As a possible application, an actuator fault diagnosis scheme is developed for single link robot arm. Finally, numerical simulation is performed to validate the effectiveness of the proposed observer in estimating the parameter along with states.