摘要: |
粒子滤波算法中重采样是解决粒子退化的一种重要方法,但重采样会导致粒子多样
性的损失。针对这一问题,对基本重采样算法进行了改进。改进算法首先按基本重采样思想
找到权值大的粒子进行复制,然后借鉴遗传算法进行交叉和变异操作,其中变异由变异尺
度因子和粒子集的均值来实现。利用改进重采样的粒子滤波算法对经典纯方位目标跟踪问题
进行了仿真,仿真结果表明,改进算法具有更好的跟踪精度。 |
关键词: 目标跟踪 粒子滤波 重采样 遗传算法 |
DOI: |
|
基金项目: |
|
An improved resampling particle filtering algorithm |
LI Shan-ji,YU Ai-lan |
(College of Engineering,Yanbian University,Yanji 133002,China) |
Abstract: |
Resampling is an important method to solve particle degradation in par
ticle filtering(PF) algorithm. But resampling will lead to the loss of particle
dive
rsity. To solve this problem, the basic resampling algorithm is improved.The imp
ro
ved algorithm first finds weights to copy large particles according to the basic
thinking
of resampling, and then uses the genetic algorithm(GA) to cross and variate.The
va
riation is realized by scale variation factor and the mean of particle sets. The
problem of classical Bearing-only target tracking is simulated with the improve
d
resampling particle filter algorithm. Simulation results show that the improved
algorithm has better tracking precision. |
Key words: target tracking particle filtering resampling genetic algorithm |