J4 ›› 2012, Vol. 39 ›› Issue (5): 154-160.doi: 10.3969/j.issn.1001-2400.2012.05.026

• 研究论文 • 上一篇    下一篇

拟蒙特卡罗聚合重采样粒子滤波无源定位算法

刘学1,2;焦淑红2;蓝晓宇2
  

  1. (1. 海军装备研究院博士后科研工作站,北京  100073;
    2. 哈尔滨工程大学 信息与通信工程学院,黑龙江 哈尔滨  150001)
  • 收稿日期:2011-06-10 出版日期:2012-10-20 发布日期:2012-12-13
  • 通讯作者: 刘学
  • 作者简介:刘学(1983-),男,哈尔滨工程大学博士研究生,E-mail: liuxue002@163.com.
  • 基金资助:

    973国家安全重大基础研究基金资助项目(61393010101-1);国家部委基础科研基金资助项目(K1503060217)

Quasi-monte-carlo merging resampling particle filter for passive location

LIU Xue1,2;JIAO Shuhong2;LAN Xiaoyu2   

  1. (1. Postdoctoral Workstation of Naval Academy of Armament,Beijing  100073, China;
    2. College of Information and Communication Engineering, Harbin Engineering Univ., Harbin  150001, China)
  • Received:2011-06-10 Online:2012-10-20 Published:2012-12-13
  • Contact: LIU Xue

摘要:

提出一种基于拟蒙特卡罗聚合重采样粒子滤波的机载无源定位算法.首先利用基于离散状态空间的粒子聚合技术对空间相近粒子进行加权聚合,在保证粒子空间分布合理性的同时有效抑制了粒子的退化;然后采用拟蒙特卡罗技术将重采样后的粒子向高似然区移动,优化了粒子在状态空间中的分布特性,提高了滤波精度.仿真结果表明: 新算法对比拟蒙特卡罗高斯粒子滤波算法,在保证滤波精度的同时,提高了运行效率.

关键词: 无源定位, 拟蒙特卡罗, 粒子滤波, 粒子聚合, 重采样, 状态估计

Abstract:

To meet the requirements of high location speed and accuracy in the air-borne passive location, a quasi-monte-carlo merging resampling particle filter is proposed. Firstly, the particle merging technique based on particles' spatial similarity is used to keep rational distribution and restrain the weight degeneracy problem of the particles. Secondly, the deterministic samples are chosen according to the quasi-monte-carlo integration technique to push the particles into the high likelihood area and explore the state space more efficiently, so that the integration error is reduced and the precision of the filter is improved. Finally, simulation results show that the efficiency of the filter is improved while the precision is quaranteed compared with the quasi-monte-carlo gaussian particle filter.

Key words: passive location, quasi-monte-carlo, particle filter, particle merging, resampling, state estimate

中图分类号: 

  • TN958.57
Baidu
map