Unknown Input Method Based Observer Synthesis and Actuator Fault Estimation for Polynomial T-S Fuzzy Systems

This paper focuses on estimating the unmeasurable states and unknown actuator fault of Polynomial Takagi-Sugeno (T-S) fuzzy system. A new approach based on the unknown input method for observer synthesis is proposed in which the unmeasurable states and the actuator faults are asymptotically estimated, simultaneously. The proposed method allows eliminating completely the effects of faults during the observer synthesis. On the basis of Lyapunov theory, the conditions for designing observer which are expressed in term of the Sum-Of Square (SOS) framework are derived. Finally, a numerical example is presented to prove the effectiveness of the proposed method.