摘要
粒子滤波（Particle Filter）是一种基于蒙特卡罗仿真的近似贝叶斯算法。其核心思想是利用一些离散随机采样点来近似系统随机变量的概率密度函数，以样本均值代替积分运算，从而获得状态的最小方差估计。对粒子滤波方法进行深入研究，进行了大量的仿真。整个仿真实验证明了粒子滤波方法在强非线性跟踪问题中的适用性和有效性。
Particle filter is a approximate Bayes estimation method based on Monte-Carol. Its nuclear idea is that some random sample is used approximating probability density function system random variable, the integral is substituted sample mean, in order to obtain minimum square error estimation of state. Particle filter is studied deeply, a lot simulation has been done. The result of simulation has shown that particle filter is practical and valid dealing with the problem of nonlinear tracking.
出处
《导航》
2009年第1期34-38,共5页
基金
国家自然科学基金资助项目（项目编号：49901013）,中国博士后基金资助项目.
