Modeling And Filtering Technology Research Based On FOG Strapdown Inertial Integrated Navigation System

With the rapid development of defense industry,it has a higher requirement for the capacity of the modern high-performance aircraft.How to improve the navigation precision of aircraft has become a hot and key issue in the field of navigation.Aiming at improving the precision of integrated navigation system for the modern high-performance aircraft,this thesis studies the relevant techniques of the fiber optical gyro(FOG)strapdown inertial integrated navigation system,which include the analysis and compensation of inertial measurement unit errors and the improvement of inertial navigation system filtering algorithm.Gyro is one of the core components of the inertial navigation system,and its output noise has a direct influence on the precision of inertial navigation system.The gyro signal de-noising methods are studied in this thesis.Firstly,a de-noising method based on the dynamic threshold of wavelet transform is proposed to improve the accuracy of FOG,which employes the components of different frequency bands decomposed by the wavelet transform to determine dynamic thresholds.Secondly,a de-nosing method based on adaptive time-frequency peak filtering(ATFPF)for FOG is proposed based on the pseudo Wigner-Ville distribution(PWVD)to reduce the noise and enhance the signal of interest.A rule for the optimal window length selection of adaptive PWVD is designed.The signal of interest is recovered by the estimation of the instantaneous frequency of the coded signal achieved by the local peak search.In addition,the methods for the analysis of the gyro random errors characteristics are studied in this thesis.An accurate analysis of the navigation system errors is achieved by analyzing the models and parameters of the gyro random errors.Firstly,the fundamentals of Allan variance analysis and dynamic Allan variance(DAVAR)analysis are studied.Secondly,a DAVAR analysis method with time-variant window lengths based on fuzzy control is proposed.According to the characteristics of FOG signal,a fuzzy controller is designed to estimate the optimum window length of DAVAR in real time.The proposed method improves the accuracy of the DAVAR results of FOG signal.Additionally,an evaluation index for the DAVAR algorithm performances is proposed based on the radar chart,which can evaluate the DAVAR algorithm performances quantitatively.Lastly,a gyro noise characteristics analysis method based on empirical mode decomposition(EMD)is proposed,which can reflect the noises distributions of different kinds of gyro in the time-frequency domain and can provide more detailed error characteristics.The access of accurate parameters of Kalman filtering algorithm,which has been commonly applied in the integrated navigation system,is beneficial for the improvement of the filtering results.Therefore the mehods for the obtainment of accurate Kalman filter parameters of the integrated navigation system are studied and a Kalman filtering algorithm based on the exact modeling for integration navigation is proposed in the thesis.Firstly,the models of gyros and accelerometers based on Allan Variance coefficients are established.Secondly,the relationships between Allan Variance coefficients and IMU parameters are derived.Lastly,error equations of inertial navigation system(INS)with the enhanced quantization noise are derived and Kalman filtering equations based on exact modeling are estabilished to obtain the accurate Kalman filter parameters.The simulation and experiment results show that the performance of the proposed algorithm is better than that of Kalman filtering.In the meanwhile,an artificial bee colony particle filter(ABCPF)algorithm for the integrated navigation system is studied in this thesis.Firstly,the fundamentals of both of the particle filter and the artificial bee colony algorithms are introduced.Secondly,aiming at the problems of particle degradation and impoverishment existed in the particle filter algorithm,an integrated navigation algorithm combining the particle filter and the artificial bee colony is proposed.The proposal distribution of particles is optimized by employing the artificial bee colony to improve the diversity of particles,and subsequently improve the accuracy of integrated navigation system.At last,a fiber strapdown inertial navigation system software platform is built based on the Matlab software in this thesis.The de-noising algorithms for FOG signal based on the dynamic threshold of wavelet transform and on the ATFPF,the algorithms for the analysis of gyro random error characteristics based on the DAVAR analysis method with time-variant window and on the EMD Allan analysis,the Kalman filtering algorithm based on exact modeling for integration navigation and the integrated navigation algorithm based on ABCPF are integrated in the developed software platform.Experiments are conducted to demonstrate the last two system filtering algorithms and it is shown that the performances of the integrated navigation system can be improved by the filtering algorithms presented in the thesis.