Research On Gyroscope Fault Detection And Diagnosis Method Based On Intelligent Information Processing

Inertial navigation system is responsible for important tasks such as navigation,positioning and attitude determination.Its accuracy and reliability have a direct impact on the safety of navigation,aviation and space.As the "brain" of inertial navigation system,gyroscope plays a pivotal role.The working state of gyroscope has a direct impact on the operation safety of the carrier,especially at present,the relationship between countries is increasingly complex,and the dependence and technical requirements on the navigation system are higher and higher.Therefore,timely access to the running state of the gyroscope and to ensure the reliability of the gyroscope is of great significance to the construction of national defense and security.This paper takes five kinds of gyroscope faults as the experimental object,and puts forward the gyroscope fault detection algorithm and fault diagnosis algorithm.The research content is mainly divided into the following steps.Firstly,the gyroscope fault data information was collected,and five kinds of fault data were modeled.Then,the gyroscope normal operation data were simulated by BP(Back Propagation)neural network.Considering that the gyroscope data usually contains some noise,and the wavelet transform has certain denoising effect.So choose wavelet analysis of BP neural network simulation the gyro fault and normal output output residuals between fault detection,wavelet packet transform to three layers of gyroscope residual signal decomposition of the energy spectrum feature fault information,and will be five kinds of fault extracting frequency band energy normalization treatment,and as a preparation for subsequent fault diagnosis of characteristic data sets.Secondly,the feature data extracted by the wavelet packet transform is taken as the input of the support vector machine,and the characteristics of the data of the gyroscope are discussed and analyzed.Finally,the radial basis function is selected as the kernel function of the support vector machine,and the one-to-one multi-classification method is adopted for classification.Several classical optimization algorithms are discussed and analyzed to optimize the relevant parameters of SVM.Finally,the cuckoo algorithm,which can be adjusted with the number of iterations and the fitness value,is selected to optimize the penalty factor and kernel parameters of SVM,so as to ensure the real-time performance and classification accuracy of fault diagnosis.Finally,LibSVM software package was used to complete the gyroscope fault detection and fault diagnosis in MATLAB,and the fault diagnosis model with better classification effect was obtained.Article through analyzing the characteristics of the gyroscope data and the characteristics and difficulties of fault diagnosis method is proposed based on BP neural network and wavelet analysis of gyro fault detection algorithm and support vector machine was optimized based on the cuckoo gyro fault diagnosis algorithm,through the experiment in the Matlab simulation,to verify the effectiveness of the two algorithms.The experimental results are analyzed and summarized.