INU fault diagnosis based on genetic wavelet neural network

This paper studies the fault diagnosis of inertia navigation unit which plays an important role in inertia navigation system. The method chosen in the fault diagnosis is combined Genetic Algorithm and wavelet neural network. Wavelet transform will effectively handle the collected inertia navigation unit signal. The characteristic signals extracted will be regarded as inputs to the neural network. The initial value of weight and bias on WNN is searched for further training by introducing Genetic Algorithm, which improves search efficiency and global convergence of the network. The fault signal of gyro which is the crucial part in inertia navigation unit is taken as an example of simulation. Simulation results indicate that this method can diagnose faults effectively.