Fault Detection Of Integrated Navigation System Based On Genetic Algorithm Optimization For Deep Belief Network

Integrated navigation system is a positioning system which can combine the navigation information of each subsystem and support each subsystem to switch each other in different working modes.It has a wide range of applications in aerospace,military and other fields.With the increasing demand for the performance of integrated navigation system,integrated navigation system has high performance,but also should ensure the good reliability of the system,so the fault diagnosis of integrated navigation system is of great significance.In the era of big data,due to the complexity of integrated navigation system and the difficulty of modeling accuracy,the data-driven intelligent fault diagnosis method is more suitable to solve the current problems.This method is more direct and effective for statistical analysis and information extraction of massive,multi-source and high-dimensional data.Therefore,the paper studies the fault diagnosis technology of integrated navigation based on GA-DBN,and realizes the fault diagnosis algorithm of integrated navigation based on GA-DBN.Aiming at the problem of data preprocessing in integrated navigation system,wavelet transform and wavelet packet decomposition reconstruction algorithm are studied.According to the characteristics of fault signal,db4 wavelet is selected as wavelet basis function to decompose the fault signal with three layers of wavelet packet,and the energy information of different frequency bands is extracted as feature vector.Aiming at the problem that the discrimination of various fault types after wavelet packet decomposition is not high enough,an improved method based on mean scale amplification is proposed,which increases the discrimination of feature vectors of different fault types,and solves the problem of the difficulty of feature vector extraction in integrated navigation system.Aiming at the problem of fault data detection in integrated navigation system,an integrated fault detection method based on Mahalanobis distance decision quantity is established.An optimized detection model based on traditional stateX~2detection and residualX~2 detection is proposed to solve the problem when the reliability of state prediction information is uncertain or the system prediction model has large errors,The over limit of detection threshold results in false warning of fault detection.Then,the optimal RBF neural network training strategy is proposed to assist the system in fault detection.Then,according to the different types of fault data detected,double adjustment robust factor and adaptive forgetting factor are established to adjust,which can effectively detect the fault data in the integrated navigation system.Aiming at the fault diagnosis of integrated navigation system,a fault diagnosis process based on deep belief network is designed.Aiming at the problem of effective training of neural network,the advantages and disadvantages of several super parameter optimization algorithms are studied.In addition,according to the characteristics of integrated navigation fault,the neural network structure and super parameter optimization method are reasonably selected.Then,a fault diagnosis algorithm based on DBN is designed and implemented,and the influence of structural parameters on the classification performance of the model is analyzed.Finally,the fault diagnosis model is constructed based on the results of parameter analysis,and the fault state of integrated navigation system is preliminarily identified.For the existing network structure and data.A fault diagnosis algorithm based on GA-DBN is designed and implemented on the basis of DBN model.Through the training and learning of historical data,we can accurately diagnose the faults of the system.To reduce the risk caused by failure.According to the processed data,it is verified simulation,and compared with the improved algorithm.It is proved that the algorithm can achieve better diagnosis effect.