Fiber Optical Gyro Fault Diagnosis based on Wavelet Transform and Neural Network

Abstract

Fault diagnosis plays an important role in detecting the reliability of integrated navigation. This paper proposed an intelligent method which combined wavelet transform with neural network to enhance efficiency. The combined method between wavelet transform and neural network was in series. Based on Daubechies, wavelet symmetry had been constructed. Through Daubechies eight-layer wavelet decomposing, detailed information of eight layers was achieved. Then the 8-dimensional eigenvector was used to train three-layer RBF neural network as fault sample. For RBF network is good at classifying, the network can detect a fault on-line after training. At the same time, it can classify faults and alarm. Gyro signals were chosen as the simulation inputs, the results indicated the method's applicability and effectiveness.