Research On Filter Algorithm Based On SINS/BDS Ultra-tight Integrated Navigation

Satellite navigation systems have a pivotal position in various fields,so at this stage all countries are striving to build satellite navigation systems.Because the performance of a single navigation system cannot meet the requirements of today's society for navigation,integrated navigation has become a popular research direction.The ultra-tight combination in the combination of the integrated navigation system is the combination with the highest degree of tightness between the systems,which effectively improves the performance of the system.Therefore,this article takes the Beidou Navigation System(BDS)/ Strap-down Inertial Navigation System(SINS)integrated navigation system as the background to carry out the research on the filtering algorithm of the ultra-tight integrated navigation system.This article designs a BDS/SINS ultra-tight integrated navigation system simulation platform based on vector tracking.It mainly includes four parts: trajectory generator design,SINS system simulator,BDS receiver simulator and integrated navigation filter.The Adaptive Unscented Particle Filter(AUPF)is proposed.In view of the strong nonlinear characteristics of the ultra-tight integrated navigation system,there will be three types of extended Kalman filter(Extended Kalman Filter,EKF),Unscented Kalman Filter(UKF),and Particle filter(PF).Non-linear algorithms are applied to the system for comparative analysis.Although the PF filtering algorithm has the best filtering effect,it has the problem of particle degradation.Unscented Particle Filter(UPF)is a more effective algorithm to solve this problem.However,the UPF algorithm still has the problem of sample impoverishment,so this article further improves the UPF algorithm and introduces an adaptive method.The algorithm first uses a self-evaluation method to test the degree of sample degradation and identify low-weight particles;secondly,a new adaptive weight adjustment method is used to migrate low-weight particles to a higher posterior probability density function.Finally,an adaptive scheme based on covariance matching technology is generated to match the actual covariance value of the residual error with its theoretical value and improve the system accuracy.The Interacting Multiple Model(IMM)algorithm is introduced again to improve the AUPF algorithm.Due to the constant changes of noise,the system estimates the noise model inaccurately.The IMM algorithm can be used to model the noise in multiple models to solve the problem of variable noise.Finally,several basic algorithms,the improved AUPF algorithm and the IMMAUPF algorithm are applied to the built simulation platform for simulation experiments.The simulation experiments show that the improved algorithm has good performance in the simulation environment.