Fault Estimation for T-S Fuzzy Systems Based on Iterative Learning Observer

The fault estimation problem for T-S fuzzy systems including actuator and sensor faults is studied in this article. The sensor fault is reconstructed by introducing a linear transformation matrix and a new measurable variable, and two observers are designed to estimate actuator and sensor faults separately. One observer is an adaptive observer that can accurately estimate actuator fault, while the other uses an iterative learning method to estimate the reconstructed sensor fault. Meanwhile, it has also been proven that the average estimation errors of the iterative learning observer converges to zero. Finally, the simulation results of numerical example validate the effectiveness of this method.