一种比例尺蒙特卡罗滤波算法

赵新; 宁小磊; 梁兴

doi:10.15892/j.cnki.djzdxb.2011.06.013

2011 , Vol. 31 >Issue 6: 179 - 182

DOI: https://doi.org/10.15892/j.cnki.djzdxb.2011.06.013

摘要

粒子滤波是一种极具潜力的非线性滤波算法,粒子退化是粒子滤波的主要问题。为此,提出一种比例尺粒子滤波算法(SPF)。在重要性抽样之后,按粒子重要性权值大小,将其分为好粒子群和较差粒子群两个种群,使用固定的或动态变化的比例尺系数对好粒子群和较差粒子群中的元素求取加权值,从而生成一些备选粒子。并通过重要性权值实现粒子的优选,从而使最好粒子参与状态估计。仿真结果表明,该算法是有效的。

关键词： 非线性滤波; 粒子滤波; 比例尺; 粒子退化

本文引用格式

赵新 , 宁小磊 , 梁兴 . 一种比例尺蒙特卡罗滤波算法[J]. 弹箭与制导学报, 2011 , 31(6) : 179 -182 . DOI: 10.15892/j.cnki.djzdxb.2011.06.013

Abstract

The particle filtering is a promising non-linear filtering, but degeneracy phenomenon is a problem. In this paper, a new particle filter (scale particle filter, SPF) was proposed to overcome the above-mentioned problem. After important sampling step, the particles were classified into two groups according to their particle-weights, and then some candidate particles were generated by the scale mean of the two groups of particles. The optimizing selection of particles was realized based on its own particle-weight so that the better particles were obtained for estimation. Simulation results demonstrate the feasibility of SPF.

Key words： non-linear filtering; particle filter; scale; degeneracy phenomenon

参考文献

[1]	Anderson BDO, Moore J B. Optimal filtering[M]. Prentice-Hall, 1979.
[2]	王法胜, 赵清杰. 一种用于解决非线性滤波问题的新型粒子滤波算法[J]. 计算机学报, 2008, 31(2): 346-352.
[3]	武元新. 对偶四元数导航算法与非线性高斯滤波研究[D]. 长沙: 国防科技大学, 2005.
[4]	De Freitas. Sequential Monte Carlo methods to train neural network models[J]. Neural Computation, 2000, 12(4): 955-993.
[5]	R van der Merwe, A Doucet, JFG de Freitas, et al. The unscented particles filter[Z]. Adv. Neural Inform. Process. Syst., Dec. 2000.
[6]	段琢华, 蔡自兴, 于金霞. 移动机器人软故障检测与补偿的自适应粒子滤波算法[J]. 中国科学 E辑, 2008, 38(4): 565-578.
[7]	宁小磊, 王宏力, 张琪, 等. 区间衍生粒子滤波器[J]. 物理学报, 2010, 59(7): 24-32.
[8]	宁小磊, 王宏力, 宁宇琪, 等. 高斯衍生粒子滤波器[J]. 西安交通大学学报, 2010, 44(6): 72-77.
[9]	Julier SJ, Uhlmann JK. Unscented filtering and nonlin-ear estimation[J]. Proceedings of the IEEE, 2004, 92(3): 401-422.
[10]	Pitt MK, Shephard N. Filtering via simulation: auxil-iary particle filters[J]. Journal of the American Statis-tical Association, 1999, 94(2): 590-599.
[11]	Jayesh H Kotecha, Petar M Djuric. Gaussian sum par-ticle filtering[J]. IEEE Transactions on Signal Process-ing, Oct. 2003, 51(10): 2602-2611.
[12]	Gordon NJ, Salmond DJ, Smith AFM. Novel approach to nonlinear/non-Gaussian Bayesian state estimation[J]. IEE Proceedings on Radar and Signal Pro-cessing, 1993, 140(2): 107-113.
[13]	Doucet A. On sequential simulation-based methods for Bayesian filtering[R]. University of Cambridge: Technical Report, CUED/F-INFENG/TR. 310, 1998.

Options

文章导航

模态框（Modal）标题

摘要

本文引用格式

Abstract

参考文献