基于势概率假设密度滤波器的不可分辨目标跟踪算法

根据有限集统计方法，推导得到了可适用于不可分辨目标跟踪问题的势概率假设密度（cardinalized probability hypothesis density， CPHD）滤波器。类似传统的点目标CPHD滤波器，该不可分辨目标CPHD滤波器不仅可以递推地传递多目标状态集合的一阶统计矩，还可以传递多目标个数（即势）的概率分布。蒙特卡罗仿真实验表明，相比Mahler提出的不可分辨目标PHD滤波器，所提出的不可分辨目标CPHD滤波器具有更加精确和稳定的多目标个数和状态估计，但它的计算量要大于不可分辨目标PHD滤波器。

Tracking unresolved targets using cardinalized probability hypothesis density filter

According to the theory of finite set statistics, a cardinalized probability hypothesis density (CPHD) filter is proposed for tracking unresolved targets. Similar to the original point-target CPHD filter, the proposed unresolved-target CPHD filter propagates not only the first-order statistical moment but also the entire probability distribution on the unresolved-target number. Monte Carlo imulation results show that the target number and state estimation from the proposed unresolved-target CPHD filter are more accurate and reliable than those of Mahler’s unresolved-target PHD filter although the computational load of the proposed CPHD filter is more expensive than that of the unresolved-target PHD filter.