Journal of South China University of Technology (Natural Science Edition) ›› 2010, Vol. 38 ›› Issue (6): 73-77,83.doi: 10.3969/j.issn.1000-565X.2010.06.014

• Electronics, Communication & Automation Technology • Previous Articles     Next Articles

Particle Filter Algorithm Based on Particle Weight Optimization in Uncertain Measurement

Hu Zhen-tao  Pan Quan  Yang Feng  Liang Yan   

  1. Institute of Control and Information,Northwestern Polytechnical University,Xi＇an 710072,Shaanxi,China
  • Received:2009-09-14 Revised:2009-10-27 Online:2010-06-25 Published:2010-06-25
  • Contact: 胡振涛（1979-），男，博士生，主要从事最优估计、非线性滤波、目标跟踪研究． E-mail:guchenshou@yahoo．com．cn
  • About author:胡振涛（1979-），男，博士生，主要从事最优估计、非线性滤波、目标跟踪研究．
  • Supported by:

    国家自然科学基金重点项目（60634030）; 国家自然科学基金资助项目（60702066）; 教育部新世纪优秀人才支持计划项目（NCET-06-0878）; 航空科学基金资助项目（20090853013）

PDF

991

Abstract:

In order to effectively measure the particle weight in uncertain measurement,a novel particle filter algorithm based on particle weight optimization is proposed.In this algorithm,first,the redundancy and complementary information among particles is fully extracted by constructing and solving the confidence distance and the confidence matrix,and a new consistency weight to measure the mutual support degree among particles is presented.Then,the weight balance factor is used to combine the cost reference weight with the consistency weight for reasonably optimizing the particle weight.The proposed algorithm not only makes full use of the information of the particles set at the current time,but also avoids the adverse effect due to the error of prior statistical information,which improves the stability and reliability of particle weight measurement.Theoretical and simulated results show that the proposed algorithm is effective.

Key words: nonlinear filtering, cost reference particle filtering, importance sampling, measurement noise