RBF neural-networks-based fault tolerant control

A fault tolerance control strategy based on RBF networks was presented. Tuning rule of the RBF networks, which guarantees the stability of the fault system was derived, developed an on-line fault tolerant control scheme using an ANN networks. The method can make the outputs of fault system tracking those of reference model without knowing the location and degree of failure, and compensating non-linear dynamics caused by the failure. The method does not need fault detection and diagnosis modular, and consequently avoid the problems of missing alarm, fault alarm and alarm delay.