中文说明：应用背景在杂波环境下，传统的多目标跟踪算法需要进行数据关联，若杂波密度或目标数较大时，关联算法变得非常复杂，会出现组会爆炸等问题。随后，Mahler提出来基于随机有限集的多目标跟踪算法，其无需进行数据关联，就可以对杂波环境下多个目标进行有效地跟踪。关键技术给出了三种基于随机有限集的多目标跟踪算法，概率假设密度（PHD）滤波器、势概率假设密度（CPHD）滤波器和势平衡多目标多伯努利（CBMeMBer）滤波器。在线性高斯条件下给出上述滤波器的高斯混合实现，并对其进行比较。
English Description：
Application backgroundIn the clutter environment, the traditional multi target tracking algorithm needs to be associated with the data. If the clutter density or the target number is large, the algorithm becomes very complex. Subsequently, Mahler proposed a multi object tracking algorithm based on the random finite set, which has no need to carry out the data association, it can track multiple targets in clutter environment.Key TechnologyThree kinds of multi object tracking algorithm based on random finite set, probability hypothesis density (PHD) filter, potential probability hypothesis density (CPHD) filter, and potential balanced multi-objective do Bernoulli (CBMeMBer) filter are presented. Gauss hybrid implementation of the filter is given under the condition of linear Gauss, and the comparison of the above filter is made.