摘要
为提高空间密集群目标分群准确率,提出一种距离划分与形状预测划分相结合的群分割算法。航迹起始阶段,通过距离划分法实现无预测信息以及预测信息不可靠条件下的有效分群;航迹维持阶段,以预测点为中心,进行目标的状态估计与形态估计,以预测形状为波门划分量测集,将落入预测波门的量测划分为一个群,并利用概率数据关联获取等效群中心、更新群航迹。经仿真验证,算法能提高群质心的估计精度,减少关联错误,提高群分割的准确性。
For space dense targets,the partition of target measurements will directly affect the performance of state estimation and filtering,so a joint partitioning algorithm combining distance partition and shape prediction partition was proposed in this paper.At the initial stage,effective clustering was achieved by distance partition under the condition of no prediction information or unreliable prediction information.In the course of track maintenance,the measurements in the predicted shape tracking gates were divided into a group,then a more precise equivalent group center was obtained by using the idea of probabilistic data association and the group center was brought into the iterative calculation of group update.The feasibility of the algorithm was verified by simulation that it can improve the accuracy of group centroid,reduce the subsequent correlation errors and improve the accuracy of tracking.
作者
韩蕾蕾
周璐
HAN Leilei;ZHOU Lu(Graduate Students Brigade of Naval Aviation University, Yantai 264001, China)
出处
《兵器装备工程学报》
CAS
北大核心
2020年第12期240-246,共7页
Journal of Ordnance Equipment Engineering
关键词
群分割
距离划分
形状预测划分
group segmentation
distance segmentation
shape prediction
作者简介
韩蕾蕾(1992—),女,硕士研究生,主要从事多目标跟踪融合研究。
