Research On Information Fusion And Error Modification Technologies For SINS/BD2 Integrated Navigation System

New optical and micro-mechanical inertial devices are gradually maturing and second-generation BeiDou Satellite Navigation System gradually completed,then on the basis of existing hardware equipments how to improve the precision and stability of the integrated navigation system,which constitute of Strapdown Inertial Navigation System(SINS)and BeiDou Satellite Navigation system(BD2),is one of the significant technologies that require to be researched in-depth.This dissertation began with error handling of strapdown inertial navigation system and nformation fusion of integrated navigation system,attaching great importance to the following aspects:As for random errors and certainty errors of inertial sensors,error filtering method by hidden markov model theory is primarily introduced and the relation between filter gain of HMM/KF(Hidden Markov Model/Kalman Filter)and denoising results of filter is analyzed to set forth extended RLS(Recursive Least Squares)method of coping with random error of inertial sensors.Then algorithm complexity is analyzed based on flops complexity analysis.Compared with HMM/KF,advantages in filter performance and filter delay and signal tracking of the proposed method are examined and analyzed.As for certainty errors,SINS position is obtained by BD2,then six positions calibration method based on rotation platform associated with extended RLS method,meanwhile extended RLS and VB(Vartional Bayesian)process noise adaptive Kalman filter associated with single-axis three positions field calibration method is proposed,Simulation results proved the efficiency of the proposed methods.The research continues on coarse and fine alignment suitable for SINS equipped with relative high precision IMU.Firstly,advantages of coarse alignment in inertial frame and attitude description by quaternion are analyzed.Then the Quest coarse alignment in inertial frame on based on quaternion attitude solution is proposed,and the advantage of this method compared with traditional Triad coarse alignment is analyzed.To satisfy the demand of fast of alignment,it is advisable to choose short time coarse alignment and nonlinear alignment method based on BD2.Then the nonlinear degree of large misalignment angles error model is analyzed.Besides,shock convergence or divergence of Cubature Kalman filter nonlinear alignment is considered due to the uncertainty of measurement model under complex field environment.To solve this problem,the AR prediction model is utilized to assist variational Bayesian adaptive filtering modifing inverse gamma parameters,and variational Bayesian Cubature Kalman filter based on AR prediction model is proposed.As to further estimating and compensating residual misalignment angle of first nonlinear alignment,second alignment based on velocity measurement and calculational angular velocity measurement,which is obtained by coupling the residual misalignment and gyro drifts is proposed.Simulation results show that the observability and precision of the proposed second alignment are enhanced.To satisfy the demand of maneuverability of navigation equipment and low cost tendency of IMU,SINS/BD2 tightly coupled navigation system under large initial error is researched.The BD2 pseudo-range and pseudo-range rate measurement equations,together with carrier phase and difference of carrier phase measurement equation are deduced.Compared the two models,pseudo-range and pseudo-range rate measurement model for tightly coupled navigation system is established.Based on the analysis of nonlinear degree of tightly coupled navigation system,the tightly coupled navigation model can be decomposed into a nonlinear model and two linear models.After that compared PF and CPF together with the proposed GSCPF(Gaussian Sum CPF),then nonlinear/linear RBCPF(Rao-Blackwellized CPF)based on choosing the CPF to cope with the nonlinear state estimation problem is proposed.The proposed RBCPF is suit for tightly coupled navigation nonlinear/linear model.Simulation results show the proposed method is superior to RBPF for tightly coupled navigation system.As for outage of BD2 satellite signal,firstly observability of the tightly coupled navigation system is analyzed.Then based on anlysis of uniform convergence of tightly coupled system on the basis of CKF,stability condition is obtained.As for partial outage of BD2,the influence of observable satellite number for navigation precision is analyzed.During BD2 outage,stability condition is dissatisfied.Digging out measurement data adequately,bridge of failure-time navigation information is realized by utilizing smoothing algorithm for the whole data navigation.Then taking posterior navigation parameters obtained from particle smoothing to replace high precision equipment as evaluation reference is proposed.For the sake of real-time accurate navigation information,IMM nonlinear filter motion constraint is introduced to restrain error accumulation of SINS during BD2 outage.As for the uncertainty of state and measurement equations introduced by the calculational motion model,the AFCKF is proposed based on the analysis that filter stability domain can be amplified by utilizing the adaptive fading factor.The stability improvement of AFCKF(Adaptive Fading Factor CKF)is proved,which ensures the stability of integrated navigation system.