该算法基本思想是利用一组带有相应权值的随机样本即粒子去逼近PHD分布和基数分布，其意义在于不仅解决重积分计算没有闭式解的难题，而且在滤波过程中，PHD函数被一系列离散的带权值的样本近似，随着样本粒子数量的增加，PHDF接近于Bayes最优估计，而且不受模型线性和高斯假设的限制，可以适用于非线性非高斯的随机系统。-The algorithm the basic idea is to use a set of random sample with the corresponding weight namely particle approximation PHD and base distribution, its significance lies in not only solve the problem of multiple integral calculation without closed solution, and in the process of filtering, PHD function by a series of discrete weighted value of sample approximation, with the increase of sample particles, PHDF close to the Bayes optimal estimation, and is not subject to the assumptions of the linear and gaussian model can be applied to nonlinear non-gaussian stochastic system.