多传感器多目标跟踪算法性能分析

Analysis of Multitarget Multisensor Tracking Algorithms

多假设与联合概率数据互联算法都能很好地解决杂波环境下的多目标跟踪问题。随后 ,人们在此基础上又将单传感器多假设、联合概率数据互联算法推广至集中式多传感器系统与分布式多传感器系统。近来 ,也有文献提出将广义S D分配算法动态化 ,实时地应用到集中式多传感器系统对杂波中的多目标进行跟踪。这些算法各有特点 ,因而它们在不同的情况下 ,算法的跟踪精度、实时性等方面各有优劣。然而 ,至今尚未见到有文献对这些算法在统一环境下进行分类比较 ,文中假定多种典型的多目标运动环境 ,对上述算法在这些环境下进行仿真实验 。

The Multitarget tracking in clutte problem can be solved weel by using multiple-hypothesis multitarget tracking algorithm and joint probabilistic data association algorithm, and these alogorithms have been applied to centralized multisensor system and distributed multisensor system. Recently, a general S-domensional assignment algorithm has been applied to centralized multisensor system to track multitarget in clutter in real time. Each of these algorithms has its own characteristics, therefore, the performance of these algorithms are different from each other in many aspects in different environments, such as tracking accuracy and real-timeness. However, no one has analyzed these algorithms in the same environment as yet. In this paper, some typical simulations are performed, and the tracking performance of these algorithms are analyzed and compared according to the results.