密集杂波环境下的快速数据关联算法

联合概率数据关联(JointProbabilisticDataAssociation,JPDA)是密集杂波环境下跟踪多目标最有效的算法之一。但当目标数目和有效量测数增大时,关联概率的计算出现组合爆炸现象一直是工程应用的瓶颈。基于JPDA算法的思想,提出了一种快速数据关联算法,该方法首先根据被跟踪目标相关门的相交情况将监视区域分成相互独立的空间,对同一空间内具有公共量测的目标和各目标相关门内的多个量测的概率密度值分别进行概率加权后再计算关联概率。不需要象最优JPDA算法中产生所有可能的联合事件,因此具有计算量小,易于工程实现的优点。仿真结果表明,在不同的杂波密度环境下和不同的目标运动形式下,此算法都可以取得令人满意的跟踪效果。

Fast Algorithm for Data Association in Dense Clutter

Joint probabilistic data association is one of the most effective algorithms of multiple target tracking in dense clutter.However,with the increment of target tracked and the number of validation measurement,the computational cost of association probability is the choke point in engineering application.A fast algorithm was developed for data association based on the principle of joint probabilistic data association.In this new method,surveillance area was divided into independent area according to the intersection area of association gate of targets tracked,after the PDF value of the common measurements of targets and measurements lying in a target association gate were weighted,and the association probability of targets were computed.Without generating all the association events,this algorithm is more efficient in engineering application.Simulation results demonstrate the tracking performance of this algorithm in different conditions.