粒子滤波的重采样方法

粒子滤波是基于递推的Monte Carlo仿真方法的总称,原则上可用于任意非线性、非高斯随机系统的状态估计.但粒子的退化现象是粒子滤波器的最大缺陷,减轻这一现象影响的方法之一就是对粒子进行重采样.阐述了几种重要的重采样算法及其改进策略,并对其进行总结与展望.

Overview of Resampling Algorithm

Particle filtering is a sequential Monte Carlo simulation algorithm, in principle, it can be used in estimating the state of any nonlinear or non-Gaussian system. But the main defect in the present particle filtering is particle degeneracy, in order to reduce the degeneracy, the resample strategy is adopted. This article introduces some resampling algorithm and their modified strategy and makes a conclusion and prospect.