The Application Of EKF Filter Combination Model In BDS Pseudo-distance Calculation

Since Kalman et al.proposed the basic principle of Kalman filtering in 1960,Kalman filtering has become an important algorithm in the field of linear valuation solution.Because most of the problems of reality are nonlinear problem,Sunahara et al in 1968 and extended kalman filter for the deformation,so it can solve the problem of nonlinear valuations,has been widely used in thefield of navigation and positioning.It has been one of the hotspots in navigation and positioning field to suppress the divergence of EKF filtering and improve the accuracy of EKF filtering estimation.The inhibition of EKF filtering divergence method is directed at the filter divergence pretreatment kalman gain to restrain filtering divergence,to the problem of EKF filtering divergence post-processing research does not see more,can’t fully guarantee the global convergence of the iterative.The EKF filter combination model adopted in this paper is an effective solution.There are two main problems in EKF filtering: initial value setting and suppression of filter divergence.Initial value is set to filter paper,"the fourth chapter 1 section combination model of EKF filtering precision of the initial value analysis" part,from two aspects of accuracy and computational efficiency of filtering the initial value is set commonly used method is analyzed.Experimental results show that compared with other initial value calculation methods,WLS method can give better consideration to the deviation and calculation efficiency of initial value.As for the methods to suppress EKF filtering divergence,chapter 3,"3.2 common methods to suppress EKF filtering divergence",is introduced in detail.Section 3.3,from the Angle of valuation reconciliation precision calculate time to adjust the gain attenuation memory method,the method,limited memory method,the prediction error covariance weighted method,adaptive method,this paper compares and analyzes the comprehensive consider valuation settlement efficiency,precision adaptive method is the best in several methods.For the post-processing method of filtering divergence,a new interest matrix combined with WLS method is introduced in the EKF filter combination model to solve the post-processing problem of filtering divergence.Judging from the new information formula,the filtering divergence is smaller and more efficient than the current estimation component mutation method.The global convergence of EKF filter combination model can be better guaranteed by starting WLS method after filtering divergence to calculate the initial value of filtering iteration and re-iterating the solution.In view of the above analysis,this article selects the WLS method,the new rate formula,the adaptive method and method of L-M EKF filter combination model and EKF filtering model portfolio selection WLS method as the initial value of the EKF filtering calculation method,can effectively reduce the EKF filter for the initial value of the calculation of deviation,which significantly improves the EKF filtering iterative speed;In the process of EKF filtering iteration,an l-m algorithm was used to modify the prediction covariance matrix of EKF filtering,which effectively ensured that the estimation error of state observation was gradually reduced and the divergence of filtering was inhibited.EKF filter combination model and to design WLS method combined with the new rate formula filtering divergence,filtering divergence WLS method is used to calculate the EKF filtering initial values,restore the EKF filtering iterative,can solve the problem of filtering divergence of post-processing,even a brief observation still missing can ensure the global convergence of EKF filter combination model.By applying EKF filter combination model to the beidou pseudo distance solution,the accuracy and applicability of the beidou-dynamic static pseudo-distance solution are analyzed.Experimental results show that the EKF filter combination model can greatly improve the accuracy of the beidou dynamic static pseudo-distance solution.In beidou pseudorange static pseudorange decoding,EKF filter combination model EKF11 compared to standard filter combination model in the iteration valuation accuracy and convergence speed 180 times before,but after iterative 180 times the accuracy of EKF filter combination model has obvious increased,the average accuracy than EKF11 model portfolio valuation standard filtering precision improved by 60%.In the calculation of the beidou pseudo-distance dynamic pseudo-distance,the EKF filter combination model improved the accuracy of dynamic point plane position coordinates by 22% and the elevation accuracy by 11% compared with the standard EKF filtering solution.In the calculation of the beidou pseudo-distance dynamic pseudo-distance,the EKF filter combination model improved the accuracy of dynamic point plane position coordinates by 22% and the elevation accuracy by 11% compared with the standard EKF filtering solution.EKF filter combination model of calculating dynamic point location,volatility is small,the location of the calculating error is smaller,more continuous excessive nature between adjacent points,closer to the GPS precise point positioning(PPP trajectory calculation of position.