基于有限集统计学理论的多目标跟踪技术具备严格的贝叶斯理论解释基础, 可以同时完成目标数目及状态的估计, 并避免了复杂的数据关联过程。基于高斯混合概率假设密度滤波器(GM-PHD)实现了对数目时变目标的跟踪, 准确估计出了每一时刻目标的数目。在此基础上进一步分析了目标生存概率 ps、目标检测概率 pd以及杂波密度 λc等因素对跟踪效果的影响, 为 GM-PHD滤波器在多目标跟踪的实际应用中各参数的取值提供了有益的参考。

The multi-target tracking techniques based on Finite Set Statistics(FISST) possess strict foundation of Bayesian theory, and it can simultaneously complete the estimations of the number of targets and their corresponding kinematic states, meanwhile the difficulties caused by traditional data association are avoided. The Gaussian Mixture Probability Hypothesis Density(GM-PHD) filter is utilized to track multi-target whose number is time-varying. The number of the targets is estimated in every moment by using this algorithm. On this basis, the influences of survival probability ps, detection probability pd and clutter density λc on the results are analyzed. The analysis results provide beneficial reference to the parameter selection of GM-PHD filter in practical applications.