密集杂波环境下用于多目标跟踪的改进PDA算法  

An Improved PDA Algorithm for Multi-target Tracking In Dense Clutter

作　　者：周红波[1] 耿伯英[1] 周荣波[2] 

机构地区：[1]海军工程大学,湖北武汉430033 [2]中国地质大学信息工程学院,北京100083

出　　处：《火力与指挥控制》2009年第3期36-38,共3页Fire Control & Command Control

基　　金：国防预研基金资助项目(10104010103)

摘　　要：概率数据关联(PDA)算法在密集波环境下跟踪单目标时具有良好的效果,但当跟踪多目标时,由于未能考虑相邻目标间的相互影响,因此效果不佳,针对PDA算法的不足,基于多目标跟踪中相邻目标间相互影响的特点,提出了一种新的改进概率数据关联(NIPDA)算法。新算法引入量测j属于目标t的概率Bjkt,并用其对关联概率进行修正,从而在计算互联概率时考虑了相邻目标间的影响。仿真结果表明,新算法在计算量与PDA算法相近的情况,提高了目标跟踪成功率。Probabilistic data association algorithm has good effectiveness when tracking single target m dense clutter. But it is not effective when tracking multi-target ,because it not consider interacting between near targets. Basing on the characteristic of interacting between near targets, a new improved PDA Algorithm is proposed in the paper. The new Algorithm import the probability of observation j belongs to target t, and modify the association probability using it. The results of simulation show that the new algorithm improves the rate of successful target tracking when its calculating quality is near to PDA algorithm.

关 键 词：多目标跟踪 数据关联 滤波 杂波 

分 类 号：TN953[电子电信—信号与信息处理]