Robust Terminal Sliding Mode Observer-Based Sensor Fault Estimation for Uncertain Nonlinear Systems

This paper presents a new scheme to design a robust terminal sliding mode observer and its application to sensor fault estimation for a class of uncertain nonlinear Lipschitz systems. In this method, first, by using transmission matrices, the nonlinear Lipschitz system is decomposed into two subsystems. Therefore, the first subsystem contains uncertainties and disturbances and the second one is constituted of fault signals. Then, using a method based on integral observer, sensor fault signal in the second subsystem is appeared as an actuator fault in the system dynamics equation. Afterwards, in order to achieve insensitivity to disturbances and ensuring a limited time convergence, a terminal sliding mode observer with H infinity performance is designed. The observer gain matrices are computed in terms of linear matrix inequality by solving an optimization problem. Then, by using equivalent output error injection concept, the reconstruction of fault could be obtained. Finally, simulation results for a single joint robot arm show the effectiveness of the proposed method.