多目标跟踪中的数据关联和航迹管理

传统的多目标关联算法JPDA存在两个缺陷:计算量较大,且其算法框架里缺乏航迹起始和终止逻辑。文中给出了一个航迹管理表,它包含多种情况下航迹的关联,能完成航迹的起始、维持和终结。在可靠航迹关联中,该文提出了一个改进的快速关联算法—多维概率数据关联(MPDA),讨论时主要考虑关联门相交区域中的公共回波对航迹更新的影响。MPDA可以类似与PDA的较低的计算代价达到了接近JPDA的关联成功率。仿真结果表明,文中的数据关联和航迹管理算法对复杂条件下的多目标跟踪能保持较高的跟踪成功率。

Data Association and Track Management in Multitarget Tracking

There are two drawbacks in the conventional multitarget association algorithm JPDA : large computation cost and lacking the logic for the track's initiation and termination in the framework.The paper develops a track management schedule,which includes several association algorithms in different circumstances.In addition,an improved fast association algorithm MPDA is proposed,which considers the effect of public measurements to the trajectories' update.MPDA can approach the association performance of JPDA with the computation cost as much as PDAF.The simulation results show that the proposed method can maintain high tracking success ratio in complicated circumstance.