CAJ | 학술논문

针对传统多目标跟踪算法在航迹初始阶段易受杂波干扰，提出一种交互多模型核预估数据流聚类的多目标跟踪算法（CE＿DMTT）。对数据流进行在线聚类，并运用交互式多模型预估类核位置，缩小聚类搜索范围，同时引入Renyi熵，对聚类进行自适应提取，获取潜在航迹。然后基于潜在航迹运用多假设跟踪算法实现实时跟踪。仿真结果表明，该算法有效减少计算复杂度，提高系统实时性。
침대전통다목표근종산법재항적초시계단역수잡파간우，제출일충교호다모형핵예고수거류취류적다목표근종산법（CE＿DMTT）。대수거류진행재선취류，병운용교호식다모형예고류핵위치，축소취류수색범위，동시인입Renyi적，대취류진행자괄응제취，획취잠재항적。연후기우잠재항적운용다가설근종산법실현실시근종。방진결과표명，해산법유효감소계산복잡도，제고계통실시성。
In order to deal with the clutter problem in the initial stage of traditional multi?target tracking,a data stream clustering mulitiple target algorithm based on Interactive Multi?Model core estimation ( CE_DMTT ) is pro?posed.Through estimating core position based on the Interactive Multi?Model to reduce the searching range of the cluster,the datastream is online clustering. In order to obtain potential track, the Renyi Entropy is introduced to adaptivly extract the cluster.Based on the obtained potential track, we use Multiple Hypotheses Tracking algorithm to implement real?time tracking.Simulation results and their analysis show preliminarily that the proposed algorithm can effectively reduce the computational complexity and improve System Real?time Performance.