基于二分图最优完备匹配的目标关联算法  

Target association method based on bipartite graph optimal complete matching

作　　者：吴汉宝[1] 李伦[1] 张志云[1] 

机构地区：[1]武汉数字工程研究所,湖北武汉430074

出　　处：《华中科技大学学报（自然科学版）》2017年第2期95-100,共6页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国防预研基金资助项目

摘　　要：针对群目标编队飞行过程中的关联问题,提出基于二分图最优完备匹配的目标关联算法.该算法利用网格邻聚构造了目标关联二分图,并给出了二分图中边的权值定义;以二分图最优完备匹配作为约束条件建立了关联模型,通过求解最优解实现了目标的正确关联.用蒙特卡罗仿真结果对所提算法在各种不同的系统偏差、目标飞行间距环境中的关联性进行了比较验证,结果表明:所提算法能够取得良好的关联效果,可以有效地抵抗传感器系统偏差的影响,同时也大大降低了密集群目标关联的不确定性,其计算复杂度能够满足实际应用需求,从而证明了该算法的有效性和鲁棒性.In order to resolve the problem of dense flying targets association,a new method based on bipartite graph optimal complete matching was proposed.In the method,the bipartite graph of target association was generated by grid clustering,and the weights of edges were defined.The target association model was introduced by the bipartite graph optimal complete matching with restrict conditions.By solving the proposed model,the target relationship can be achieved properly.The proposed association method was verified by Monte-Carlo simulation experiment in different systematic errors and targets flying interval situation.Experimental results indicate that the algorithm can restrain the influence of the sensor system errors,reduce the uncertainty of denseness multi-target association problem,and satisfy the computational complexity of data fusion engineering application,which show good performance with validity and robustness.

关 键 词：群目标 系统偏差 二分图 目标关联 最优匹配 

分 类 号：TN957[电子电信—信号与信息处理] TP274[电子电信—信息与通信工程]