Fault diagnosis observer via hybrid fuzzy and importance sampling schemes

This paper introduces a new direction of research to estimate states as well as detect and diagnose (Gaussian and non Gaussian) faults. Therefore, a new observer (FO) has been introduced and designed via a new filter for each output of plant. The new filter FSISF based on Fuzzy and Sequential Important Sampling algorithms to estimate and predicates the. Furthermore, the observer estimates the unknown states of the plant according to the diagnosed fault, previous predicate weight and the residual of the plant. As results, a nonlinear Dc motor model considered as a benchmark to test the new observer (FO), where, the good results of the simulation results have shown that the proposed observer is a robust observer against the colored, white noise and non Gaussian noise and fault.