粒子滤波(PF: Particle Filter)的思想基于蒙特卡洛方法(Monte Carlo methods)，它是利用粒子集来表示概率，可以用在任何形式的状态空间模型上。其核心思想是通过从后验概率中抽取的随机状态粒子来表达其分布，是一种顺序重要性采样法(Sequential Importance Sampling)。简单来说，粒子滤波法是指通过寻找一组在状态空间传播的随机样本对概率密度函数 进行近似，以样本均值代替积分运算，从而获得状态最小方差分布的过程。这里的样本即指粒子,当样本数量N→∝时可以逼近任何形式的概率密度分布。-Particle filter (PF: Particle Filter) ideas based on Monte Carlo methods (Monte Carlo methods), which is set to represent the probability of a particle, can be used in any form of state space model. The core idea is to extract from the posterior probability of the random state of particle to express the distribution is a sequential importance sampling method (Sequential Importance Sampling). In short, particle filtering method is by looking for a spread in state space probability density function of random samples to approximate to the sample mean instead of integral operators to gain distribution in the state minimum variance process. Here' s the sample i.e. particles, when the sample size N → α can approach any form of probability density distribution.