多目标跟踪中几种数据关联方法的比较

数据关联是多目标跟踪的核心部分,概率数据关联滤波（probability data association filter, PDA）算法和联合概率数据关联滤波（joint probability data association filter,JPDA）算法是两种典型的数据关联算法.PDA方法计算量小,有较好的单目标跟踪性能,但在跟踪多个目标时会发生目标的偏移和聚合现象;JPDA方法具有良好的多目标跟踪性能,但其运算量巨大而难于在工程中应用.OSPDA（the order statistics PDA）是改进型的快速算法,关联精度高并且计算量小.对这几种数据关联方法进行了分析比较,并作了仿真,指出各种算法的优缺点.

The Comparison of Data Association Algorithms for Multitarget Tracking

Data association is the core of multitarget tracking. The probability data association filter (PDA) and the joint probability data association filter (JPDA) are the two typical algorithms. The PDA algorithm has low computational burden and is a good scheme for single target tracking, but it can't solve the track coalescence in multitarget tracking. The JPDA algorithm has great capability in multitarget tracking but can't be easily implemented in reality because of very high computational load. OSPDA (the order statistics PDA) is an improvement of data association. The computational burden of OSPDA is low and the tracking accuracy of OSPDA is high. In this paper, the comparison of the three methods is proposed and through computer simulation the advantages and disadvantages of the three methods are proposed.