A new robust observer design for nonlinear systems with application to fault diagnosis

This paper presents a robust fault diagnosis scheme for a class of uncertain nonlinear systems whose nonlinear function satisfies the Lipschitz condition with unmatched time-varying uncertainties, external disturbances and perturbed output. The design procedure combines the high robustness of the nonlinear unknown input observer with sliding-mode techniques in order to enhance the estimation qualities. The proposed design is derived and expressed as a linear matrix inequality optimization problem. Additionally, we have provided an approach to reduce conservatism in the derivation of the stability conditions. The effectiveness of this observer and the fault diagnosis scheme are shown by applying them to a single-link manipulator. Simulation results are presented to validate the proposed approach and show the robustness for the system nonlinearity and unmatched time-varying uncertainties.