西安电子科技大学学报

• 研究论文 • 上一篇    下一篇

估计偏差修正扩展卡尔曼滤波新算法

邓兵;孙正波;贺青   

  1. (盲信号处理重点实验室,四川 成都 610041)
  • 收稿日期:2017-04-17 出版日期:2018-06-20 发布日期:2018-07-18
  • 作者简介:邓兵(1987-),男,博士,E-mail: dengbing0317@sina.com
  • 基金资助:

    国家自然科学基金资助项目(61172140,61304264)

Novel extended Kalman filter with linear-correction

DENG Bing;SUN Zhengbo;HE Qing   

  1. (National Key Lab. of Science and Technology on Blind Signal Processing, Chengdu 610041, China)
  • Received:2017-04-17 Online:2018-06-20 Published:2018-07-18

摘要:

针对扩展卡尔曼滤波跟踪器在非线性系统目标跟踪过程中容易发散这一难题,在时差频差测量体制下,提出了一种基于估计偏差修正的扩展卡尔曼滤波算法.首先利用加权最小二乘算法求解标准扩展卡尔曼滤波状态估计偏差并进行线性修正;然后重新计算观测矩阵并且再次进行滤波估计,以减小局部线性化截断误差对于观测矩阵的影响,提升目标状态跟踪精度.仿真结果证明了所提算法具有较好的目标跟踪性能.

关键词: 目标跟踪, 扩展卡尔曼滤波, 到达时差, 到达频差, 偏差修正

Abstract:

This paper considers the divergence problem of the extended Kalman filter(EKF) in performing estimation of the state of a nonlinear system. A linear-correction EKF algorithm that involves closed-form weighted least squares(WLS) optimization only is developed using the time difference of arrival(TDOA) and frequency difference of arrival(FDOA) measurements The proposed method first estimates and subtracts the initial state estimation error in the result of EKF using the WLS approach, then recycles the filtering process again to overcome the error caused by the part linearization operation for the measurement matrix in EKF. Simulation results demonstrate the good performance of the proposed method.

Key words: target tracking, extended Kalman filter, time difference of arrival, frequency difference of arrival, linear-correction