To address the problem that the output signal of the gyroscope in an inertial navigation system is nonlinear and that the fault characteristics of this signal are not obvious and improve the correct rate of fault diagnosis accuracy of inertial devices in inertial navigation systems,a gyroscope fault diagnosis method based on improved particle swarm optimization algorithm(PSO)and probabilistic neural network(PNN)is proposed.Firstly,for the four common fault signals during the operation of gyroscope,mathematical models are established and fault feature coefficients are extracted by wavelet transform.Secondly,the particle update of particle swarm uses Cubic chaos mapping and nonlinear decreasing inertia weight coefficients in a way,and the improved particle swarm optimization algorithm is used for optimal smoothing factor selection of probabilistic neural networks.Finally,the probabilistic neural network is trained to classify and diagnose the gyroscope fault signals.The offline test results show that the CPSO-PNN network achieves an average correct rate of 95.8%for the four fault classifications.
NETL
NSTL
万方数据
