Actuator fault diagnosis based fuzzy multiple model structure for moving systems

In this paper, the fuzzy multiple model scheme is proposed based on a suitable combination of Takagi-Sugeno(T-S) fuzzy model and multiple model technique for actuator failure. The non-linear system is described by T-S fuzzy model. Based on multiple model structure for locally linearized models, an adaptive observer approach is studied. The fuzzy adaptive observer is described by a fuzzy fusion of all locally adaptive observers. We developed an adaptive approach to estimate the unknown parameters on-line, so that the fault can be found when the fuzzy adaptive observer model matches the system and the lock-in-place fault can be estimate accurately. Numerical simulation results on an aircraft example show satisfactory performance of the proposed algorithm.