SIGMA-point Non-linear Filter And Its Application Research

This paper mainly studies the Sigma-point non-linear kalman filter(SPKF)and designs an atillery locating simulation system.Firstly,the sampling rules and detailed algorithms of the classcal SPKF of UKF,CDKF and CKF are introduced,then a NISRCKF algorithm is proposed,which inherites the advantages of fast and robust of SRCKF,combines the Gauss-Newton iterative theory and the noise compensation method to update the process of the measurement of SRCKF,it makes full use of the latest measurement information and the filtering effect can be improved by selecting the appropriate noise compensation factor.The advantages of the other SPKF are reflected in the simulation experiment of reentry trajectory target state,and the filtering accuracy is improved with the increase of the iteration numbers.Secondly,aiming at the operational efficiency of nonlinear filtering algorithm,a square-root UCKF filtering algorithm with scaled minimal skew simplex sampling(SSRUCKF)is proposed.In the prediction,the scaled minimal skew simplex sampling strategy is introduced into the UT transformation of SRUKF,the non-linear transformation is carried out with fewer Sigma-points.In the measurement,the Spherical-Radial Cubature rule is used to select Sigma-points and transmit the square-root of the covariance.Comparing with other SPKF,SSRUCKF use less Sigma-points in the prediction,improves the operation efficiency while maintaining excellent filtering accuracy,which is suitable for complex non-linear systems with higher real-time requirements.Finally,this paper analyzes the dynamic model of the artillery,designs a simulation system software for atillery locating,and verifies the practical performance of non-linear filter algorithms,then a series of experiments show availability of the system.