Fuzzy Augmented State Kalman Observer For Fault And State Estimation

The paper studies the problem of simultaneously state and fault estimation of the non linear stochastic time varying system. An approach based on fuzzy augmented state kalman observer (FASKO) is developed to solve the problem stated above. It consist of combining fuzzy Takagi Sugeno dynamic model with the kalman filter theory. At each local linear model of the fuzzy model, Kalman filter equations are used. The performance of the FASKO have been compared to the classical augmented state kalman filter. Simulation results performed on three tank system, illustrate the effeciency of the proposed approach.