Fuzzy neural networks is applied in diagnosing the fault of control system

This paper set up monitoring management of fuzzy neural networks based on number type, semantic type and indefinite information character through the combination of fuzzy classification and neural networks. If general neural networks is given to fuzzy input signal or fuzzy authority value, fuzzy neural networks, which is the system including language computation, logic inference, distributed processing and nonlinear whose performance are just satisfied with processing demand of multi-sensor information fusion mapping, show complementation of fuzzy system and neural networks. Finally, by means of fault diagnosis experiment of attitude control system, we compare approximation performance of fuzzy neural network with BP neural network; approximation effect and stable performance of FNN is equivalent to BPNN, but convergence rate of the former is much faster than the latter.