一种扩展H_∞粒子滤波方法  

A Extended H_∞ Particle Filter Algorithm

作　　者：万洋[1] 王首勇[2] 于兴伟[1] 

机构地区：[1]空军雷达学院研究生管理大队,武汉430019 [2]空军雷达学院重点实验室,武汉430019

出　　处：《信号处理》2010年第6期869-874,共6页Journal of Signal Processing

摘　　要：重要性函数的选择是粒子滤波算法的核心,本文提出一种基于扩展H∞滤波(EHF)产生重要性函数的扩展H∞粒子滤波(EHPF)算法,由于EHF滤波算法鲁棒性强、滤波精度高,且该滤波算法考虑了最新的观测数据,因此由其产生的重要性函数更接近于系统状态的真实后验概率分布。理论分析和仿真结果表明扩展H∞粒子滤波算法的滤波性能明显优于标准粒子滤波算法,扩展卡尔曼滤波算法和扩展卡尔曼粒子滤波算法,与不敏粒子滤波算法滤波精度相当,但计算复杂度要低于不敏粒子滤波算法,是一种有效的粒子滤波算法。The choose of important function is a critical issue in particle filter algorithm,in the paper we propose a extended H∞ particle filter（EHPF） algorithm with a important function generated by the extended H∞ filter（EHF）.Because the extended H∞ filter algorithm has very high accuracy and strong robustness,and the filter algorithm integrates the new observations,then the important function which it generates can approximate the real posterior probability distribution of the system state reasonable well.The theoretical analysis and experimental results show that the extended H∞ particle filter algorithm is superior to the standard particle filter algorithm and others filters algorithm such as the extended kalman filter algorithm and extended kalman particle filter algorithm,provides performance comparable to that of the unscented kalman particle filter algorithm but with lower computational cost,so it＇s a effective particle filter algorithm.

关 键 词：粒子滤波 重要性函数 扩展H∞粒子滤波 

分 类 号：TN953[电子电信—信号与信息处理]