递推更新高斯粒子滤波器

发布时间：2018-05-21 03:10

本文选题：高斯滤波器 + 粒子滤波　；参考：《控制理论与应用》2016年03期

【摘要】：传统高斯粒子滤波算法(Gaussian particle Filter,GPF)中,粒子的重要性密度函数是由高斯滤波器结合当前最新量测来构建的.由于传统高斯滤波器在量测更新阶段直接利用量测对状态进行线性更新,在某些条件下会导致所构建的重要性密度函数并不能很好地近似状态真实分布.为了解决这一问题,结合递推更新的思想,本文推导出了递推更新高斯滤波器(recursive update Gaussian filter,RUGF)的一般结构.并在此基础上,选用RUGF来构建粒子滤波的重要性密度函数,从而提出了基于递推更新的高斯粒子滤波算法(recursive update gaussian particle filter,RUGPF).仿真表明,在非线性系统状态估计问题中,递推更新可以很好的利用量测信息,相比于传统的GPF,本文所提出的RUGPF滤波算法可以提供更高精度的估计结果.
[Abstract]:In the traditional Gauss particle Filter (GPF) algorithm (Gaussian Filter, GPF), the importance density function of the particle is constructed by the Gauss filter combined with the latest measurement. Because the traditional Gauss filter uses the measurement to update the state directly in the measurement update stage, it will lead to the importance of the construction under some conditions. In order to solve this problem, in order to solve this problem, the general structure of recursive update Gaussian filter (RUGF) is derived in this paper, and based on this, the importance density function of particle filter is constructed by RUGF, and the basis is proposed. The recursive updated Gauss particle filter algorithm (recursive update Gaussian particle filter, RUGPF). The simulation shows that in the nonlinear system state estimation problem, the recursive update can make good use of the measurement information. Compared to the traditional GPF, the proposed RUGPF filter algorithm can provide more accurate estimation results.
【作者单位】：哈尔滨工程大学自动化学院;
【基金】：国家自然科学基金项目(61201409,61371173) 中国博士后科学基金(2013M530147,2014T70309) 黑龙江省博士后基金(LBH-Z13052,LBHTZ0505) 哈尔滨工程大学中央高校基本科研业务费专项基金(HEUCFQ20150407)资助~~
【分类号】：TN713

【相似文献】