粒子滤波算法原理及其实验性能分析

Algorithm Principle of Particle Filter and Analysis of Performance

在介绍粒子滤波的基本思想的基础上,给出了粒子滤波算法的基本原理和实现步骤,并结合扩展卡尔曼滤波算法,在实验环境下进行了包括粒子数、重采样方法等在内的性能分析,得出了一些重要结论。结果表明,处理非线性系统时,粒子滤波精度要高于EKF;粒子数与算法精度和运算复杂度有直接关系;粒子滤波可以处理多维系统状态。

The basic principle and the implementation steps of particle filter are given based on the basic idea in this paper. The performance including the number of particles and resample algorithm is analyzed combined with EKF and the experimental results show some conclusions that the particle filter is more accurate than EKF on the treatment of nonlinear systems; the precision and complexity of algorithm is related to particle number; particle filter can process multidimensional system.