基于随机有限集滤波器的可分辨群目标跟踪技术研究综述  

RFS-Filters-Based Resolvable Group Target Tracking Technology:A Review

作　　者：杨超群 徐梦蝶 梁潇洧 朱鑫潮 张恒 曹向辉[1] YANG Chaoqun;XU Mengdie;LIANG Xiaowei;ZHU Xinchao;ZHANG Heng;CAO Xianghui(School of Automation,Southeast University,Nanjing,Jiangsu 210096,China;College of Information Science and Electronic Engineering,Zhejiang University,Hangzhou,Zhejiang 310030,China;School of Computer Engineering,Jiangsu Ocean University,Lianyungang,Jiangsu 222005,China)

机构地区：[1]东南大学自动化学院,江苏南京210096 [2]浙江大学信息与电子工程学院,浙江杭州310030 [3]江苏海洋大学计算机工程学院,江苏连云港222005

出　　处：《信号处理》2024年第10期1763-1772,共10页Journal of Signal Processing

基　　金：江苏省自然科学基金(BK20230827);国家自然科学基金(62303109);东南大学至善青年学者支持计划(2242024RCB0011);东南大学新进教师科研启动经费(RF1028623002)。

摘　　要：群目标跟踪在救灾搜救、海陆防御和战场作战等领域展现出广阔的应用前景。与传统多目标跟踪不同,群目标跟踪不仅涉及对多个个体目标的跟踪,还涉及对一个群体目标的跟踪。在群体中各个子目标需要同步运动以避免碰撞,同时群体中的子目标数量和群结构还将随着时间推移而改变。根据子目标数量和传感器分辨率的不同,群目标跟踪问题包含可分辨、不可分辨、部分可分辨以及部分不可分辨群目标跟踪等多个类别。其中,可分辨群目标跟踪问题需要同时对群体结构、群内子目标交互和数目进行估计。现有研究主要关注基于传统数据关联和随机有限集滤波器的可分辨群目标跟踪方法,其中,基于随机有限集滤波器的方法通过将多个目标状态联合建模成随机有限集,缓解了数据关联问题,从而可更好地适应跟踪场景。为更好地展示群目标跟踪方法的研究进展,综述了近年来基于随机有限集滤波器的可分辨群目标跟踪的若干代表性方法,包括基于多目标多伯努利滤波器、基于标签随机有限集滤波器和基于泊松多伯努利混合滤波器的群目标跟踪方法。这些方法在处理可分辨群目标跟踪问题时展示出了显著的优势。最后,探讨了存在的问题和未来的发展方向。Group target tracking has demonstrated broad application prospects in disaster relief and search and rescue,sea and land defense,and battlefield operations.Unlike the conventional multitarget tracking,group target tracking involves the tracking of not only multiple individual targets but also that of a group target.In a group,each subtarget must propagate synchronously to avoid collision;additionally,the number of subtargets and the structure in the group change over time.Depending on the number of subtargets and the sensor resolution,group target tracking can be classified into resolvable,indistinguishable,partially resolvable,and partially indistinguishable group target tracking.Hence,the problem associated with resolvable group target tracking requires the simultaneous estimation of the group structure as well as the interaction and number of subtargets within the group.Existing studies focus primarily on resolvable grouptarget-tracking methods based on the conventional data association and random-finite-set filters.Among them,the method based on random-finite-set filters alleviates the data-association problem by jointly modeling multiple target states as random finite sets,thus enabling better adaptation to tracking scenarios.To illustrate the research progress of resolvable group-target-tracking methods more clearly,some representative methods based on random-finite-set filters proposed in recent years are reviewed,including methods based on multitarget multi-Bernoulli filters,labeled randomfinite-set filters,and Poisson multi-Bernoulli mixture filters.These methods are particularly advantageous for solving problems associated with resolvable group target tracking.Finally,the existing problems and future directions are prospected.

关 键 词：群目标跟踪 可分辨群目标跟踪 随机有限集 多伯努利滤波器 

分 类 号：TP14[自动化与计算机技术—控制理论与控制工程]