Abstract: In view of the defects when extended Kalman Filter is employed in the nonlinear land vehicle system, this paper introduces a new method—Sigma Point Kalman Filter in the navigation state estimation problem. The idea of the filter is based on weighted statistical linear regression technology, the SPKF can generate better estimates of mean and the covariance of the state. Compared with the filtering estimate results given from a time series state estimation example based on GPS/DR Model navigation sysytem, the simulation shows that SPKF not only leads to more accurate results than EKF, but also avoid computing Jacobi matrix. So it is an ideal nonlinear filtering method in the land vehicle navigation.

Key words: Kalman Filter, State estimation, Sigma points

摘要： 针对扩展卡尔曼滤波(EKF)在车载组合导航系统状态估计问题中的缺陷，本文介绍了一种新的方法——Sigma点卡尔曼滤波(SPKF)用于车载组合导航系统的非线性状态估计。其思想是基于非线性函数的加权统计线形化,SPKF滤波算法能够给出随机变量非线性变换以后更精确的均值和协方差的估计,从而带来更高的精度。最后通过GPS/DR组合导航模型时间序列的状态估计仿真实例说明:同EKF相比,SPKF的滤波精度和稳定性都显著提高了,还可避免计算烦琐的Jacobi矩阵,是一种良好的非线形滤波方法。

关键词: 卡尔曼滤波, 状态估计, 西格马点