无线电工程2025，Vol.55Issue(6)：1335-1341,7.DOI:10.3969/j.issn.1003-3106.2025.06.022
基于模糊估计和最大权值匹配的多目标跟踪算法
Multi-target Tracking Algorithm Based on Fuzzy Estimation and Optimal Weight Matching
摘要
Abstract
Plot-track association algorithm is the core of radar multi-target tracking,and traditional data association algorithms,such as Nearest Neighbor(NN)Data Association and Joint Probabilistic Data Association(JPDA),have problems such as simple association logic or complex calculation.As a result,in complex scenarios,radar systems often fail to realize target plot-track association,leading to loss of tracking.Additionally,the real-time performance is often too poor for direct deployment in real-world products.A multi-target tracking algorithm based on fuzzy estimation and optimal weight matching is proposed.The algorithm first starts all tracks in the current situation during the tracking process.Then,the detected measurement plots and all the tracks in the situation are fuzzily matched and an association weight matrix is established.Finally,with the objective of maximizing the total association weight of the matrix,the Kuhn-Munkres algorithm is used to obtain the optimal matching measurement plots of all tracks that maximizes the global weight from the association weight matrix.Finally,the simulation is compared with NN and JPDA,and the proposed algorithm is applied in the measured data.Experimental results show that the proposed algorithm not only solves the problem of tracking error caused by association error,but can also avoid the influence of clutter and other tracks in practical applications and maintain stable tracking of targets.Moreover,its computational load remains within acceptable limits,demonstrating strong potential for engineering deployment.关键词
多目标跟踪/点航关联/模糊估计/Kuhn-Munkres算法Key words
multi-target tracking/plot-track association/fuzzy estimation/Kuhn-Munkres algorithm分类
信息技术与安全科学引用本文复制引用
薛俊杰,刘良玉..基于模糊估计和最大权值匹配的多目标跟踪算法[J].无线电工程,2025,55(6):1335-1341,7.基金项目
国家部委基金资助项目 Project Funded by National Ministries and Commissions （）
