两类典型多目标跟踪算法的性能分析与比较  

Performance Analysis and Comparison of Two Classic Algorithms in Multi-target Tracking

作　　者：王芝[1] 徐晓滨[1] 刘伟峰[1] 文成林[1] 

机构地区：[1]杭州电子科技大学信息与控制研究所,杭州310018

出　　处：《电子与信息学报》2010年第7期1633-1637,共5页Journal of Electronics & Information Technology

基　　金：国家自然科学基金(60772006;60934009);浙江省自然科学基金(R106745;Y1080422);清华信息科学与技术国家实验室学科交叉基金等课题资助

摘　　要：在处理目标跟踪的两类主要方法中,一类是通过数据关联来解决,如PDA和JPDA等;另一类则是绕过关联直接处理,如随机集、GM-PHD等。该文从两类典型方法中各选取一种有代表性的方法,如JPDA与GM-PHD,首先通过分析两种算法主要步骤的计算量,得到相应算法总计算量的解析表达式;然后根据观测与目标状态之间关联复杂程度,分3种情况对两类算法的计算量进行比较;最后以仿真说明算法的跟踪效果,并以算法运行时间来验证计算量公式的正确性。There are two primary ways to process multi-target tracking problem.One is data association method,whose deputies are PDA and JPDA.The other is direct method without the data association,whose deputies are random sets theory and GM-PHD.Two representational algorithms are chosen from aforementioned two kinds of methods respectively,that is,JPDA and GM-PHD.Firstly,general analytical forms to evaluate calculation complexity of each algorithm are formulated by analyzing and totaling their major operation steps.Secondly,the calculation complexity of two algorithms is compared through three cases respectively,which are divided on the basis of associated complexity between states and the measurements.Finally,one example,including tracking effect and the running time,is utilized to illustrate the analytical forms of evaluating calculation complexity proposed in this paper.

关 键 词：多目标跟踪 联合概率数据关联(JPDA) 随机集 概率假设密度(PHD) 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]