计算机工程 ›› 2012, Vol. 38 ›› Issue (08): 177-179.doi: 10.3969/j.issn.1000-3428.2012.08.058

• 人工智能及识别技术 • 上一篇    下一篇

多观测信息动态加权粒子滤波

白剑锋,南建国,邬 蒙,查 翔   

  1. (空军工程大学工程学院,西安 710038)
  • 收稿日期:2011-09-02 出版日期:2012-04-20 发布日期:2012-04-20
  • 作者简介:白剑锋(1985-),男,硕士研究生,主研方向:信息融合,目标跟踪;南建国,副教授;邬 蒙,讲师;查 翔,硕士研究生

Dynamic Weighted of Multiple Observation Information for Particle Filtering

BAI Jian-feng, NAN Jian-guo, WU Meng, ZHA Xiang   

  1. (Engineering College, Air Force Engineering University, Xi’an 710038, China)
  • Received:2011-09-02 Online:2012-04-20 Published:2012-04-20

摘要: 在机动目标跟踪中,当系统状态向量为多维的情况下,单一观测量的滤波跟踪无法满足多维状态估计精度的要求。为此,提出一种基于粒子滤波融合多观测量的动态加权算法。该算法利用多个高度非线性观测量,并通过动态加权方法融合多个估计值,提高机动目标跟踪的精度。仿真实验验证了该算法的有效性。

关键词: 多观测量, 动态加权, 粒子滤波, 机动目标跟踪, 径向速度, 蒙特卡罗方法

Abstract: It can not satisfy the accurate need of the estimation of the multiple dimension state vector by using a single observational value. So more kinematic observational information must be used for the calculation. On the basis of this, this algorithm based on dynamic weighted multiple observational values for particle filtering is presented. This algorithm uses multiple observational values and can enhance the precision of target tracking by combining multiple estimated state vectors. Simulation experiment verifies the effectiveness of the algorithm.

Key words: multiple observational values, dynamic weighted, particle filtering, maneuvering target tracking, radial speed, Monte Carlo method

中图分类号: