一种改进的联合概率数据关联算法

Modified Joint Probabilistic Data Association Algorithm

对于多目标跟踪问题,数据关联是其核心部分,联合概率数据关联算法(JPDA)是多目标跟踪的典型方法。当目标较为密集,计算量剧增,会出现计算组合爆炸现象,而其本质就在于确认矩阵拆分成可行矩阵的计算量。为了降低JPDA的计算量,本文提出了一种改进的JPDA算法,在拆分确认矩阵时引入分支定界算法的思想,以确定每个目标的最后一个回波。当搜索到最后一个回波时停止搜索,执行下一个目标回波的搜索,直至结束。利用该改进算法对杂波环境下多目标跟踪进行仿真实验,结果表明,该算法使其时间代价减少。

The data association is the core of multi-target tracking and joint probabilistic data association algorithm（JPDA） is a typical method for multi-target tracking.But when the targets become very dense,the computational load is higher to cause the combinational explosion problem.The basic reason for the problem is that the computation of the confirmed matrix is divided into feasible matrices.To reduce the computation of JPDA,a modified algorithm of JPDA is proposed,which is inspired by the principle of Bound and Branch algorithm,and the last echo of each target can be determined.Stop searching when the echo is the last,and search next echo of target,until the end.Simulation results show that the time complexity is reduced by using the proposed algorithm for multi-target tracking in the clutter.