结合总体最小二乘的自适应抗差滤波算法  

Adaptively Robust Filtering Algorithm Combined with Total Least Squares

作　　者：刘青[1] 崔先强[1] 宋迎春[1] 

机构地区：[1]中南大学地球科学与信息物理学院,湖南长沙410083

出　　处：《科技视界》2015年第24期64-66,97,共4页Science & Technology Vision

基　　金：地理信息工程国家重点实验室开放基金(SKLGIE 2013-M-2-4;SKLGIE 2013-M-2-5);中国博士后科学基金(2013M540641)

摘　　要：在动态导航定位中,不仅会存在有色噪声的影响,而且还会出现观测系数矩阵存在偏差的情况,数据处理中不容忽视。针对此,本文将总体最小二乘和自适应抗差滤波相结合,提出了能够同时控制有色噪声和系数矩阵偏差影响的新方法。该算法先用总体最小二乘方法求出观测方程系数矩阵偏差,并对观测方程系数矩阵进行修正,再基于修正后的观测系数矩阵使用自适应抗差滤波算法求解出最后的状态参数估值。计算结果表明,该方法不仅能够改正系数矩阵偏差且能控制有色噪声的影响,有效地提高导航定位的精度和可靠性。In the kinematic navigation and positioning, not only there are the influence of colored noise, but also the observation coefficient matrix can appear the deviation. These errors can not be ignored in data processing. For this, a new method combined total least squares with adaptively robust filtering is presented to control the influences of the colored noises and the coefficient matrix deviation at the same time. The algorithm first use Total Least Squares to obtain the coefficient matrix deviation of observation equation, and then use this deviation to correct the observation equation coefficient matrix. Further adaptively robust filtering algorithm will be used to solve the final state parameter estimation. Finally,the feasibility and effectiveness of the algorithm is verified by navigation data. The results show that this algorithm can correct the coefficient matrix deviation and control the influence of colored noise effectively.

关 键 词：总体最小二乘 卡尔曼滤波 自适应抗差滤波 动态导航 

分 类 号：TN911.7[电子电信—通信与信息系统]