非线性系统中状态和参数联合估计的双重粒子滤波方法

侯代文; 殷福亮

doi:10.3724/SP.J.1146.2007.00273

非线性系统中状态和参数联合估计的双重粒子滤波方法

doi: 10.3724/SP.J.1146.2007.00273

侯代文^{①② 殷福亮,},
殷福亮

基金项目:

国家自然科学基金(60772161，60372082)和教育部跨世纪优秀人才基金资助课题

计量
- 文章访问数: 3965
- HTML全文浏览量: 106
- PDF下载量: 1272
- 被引次数: 0
出版历程
- 收稿日期: 2007-02-13
- 修回日期: 2007-09-28
- 刊出日期: 2008-09-19

A Dual Particle Filter for State and Parameter Estimation in Nonlinear System

Hou Dai-Wen^{①② 殷福亮
,},
Yin Fu-Liang

摘要

摘要: 该文提出了一种双重粒子滤波方法，对存在未知参数的非线性系统进行状态和参数联合估计。该方法采用基于充分统计量的粒子滤波技术，避免了重采样过程中的粒子枯竭现象；采用贝塔分布拟合系统参数的后验分布，不仅充分利用了先验信息，而且避免了对高斯分布拖尾部分的采样，提高了粒子的采样效率。仿真实验结果表明，该方法提高了非线性系统中状态和参数的估计精度，降低了滤波器对初始误差的敏感性。
- 粒子滤波 /
- 双重估计 /
- 充分统计量 /
- 贝塔分布 /
- 非线性系统
Abstract: The dual particle filter is proposed to solve the problem of simultaneously estimating the state and the parameter of a nonlinear dynamic system. In the new filter, the sufficient statistics based particle filter is adopted to deal with sampling impoverishment arising in generic particle filter and the beta distribution, which makes good use of the prior knowledge as well as avoids tail draws for the parameter, is used to fit the parametric a posteriori probability density function. Simulation results show that both estimation accuracy and initial sensitivity of the nonlinear system are improved.
- P /
- a /
- r /
- t

HTML全文

参考文献(1)

[1] Kalman R. A new approach to linear filtering and predictionproblem. Trans. of the ASME-Journal of Basic Engineering,1960, 82(D): 34-45. [2] Jazwinski A H. Stochastic Processes and Filtering Theory.New York: Academic press, 1970: 281-286. [3] Dempster A P, Laird N M, and Rubin D B. Maximumlikelihood from incomplete data via the EM algorithm.Journal of the Royal Statistical Society, 1977, Series B, 39(1):1-38. [4] Goodwin G C and Agero J C. Approximate EM algorithmsfor parameter and state estimation in nonlinear stochasticmodels. Proceedings of the 44th IEEE Conference onDecision and Control, and the European Control Conference2005. Seville, Spain, 2005: 368-373. [5] Lange K A. Gradient algorithm locally equivalent to the EMalgorithm. Journal of the Royal Statistical Society, 1995,Series B, 59(2): 425-437. [6] Berzuini C and Best N G, et al.. Dynamic conditionalindependence models and Markov chain Monte Carlomethods[J].Journal of the American Statistical Association.1997, 92(440):1403-1441 [7] Gordon N, Salmond D, and Smith A F M. Novel approach tononlinear and non-Gaussian Bayesian state estimation. IEEProceedings-F, 1993, 140(2): 107-113. [8] Liu J and West M. Combined parameter and state estimationin simulation-based filtering. in Sequential Monte Carlo inPractice, A. Doucet, N. de Freitas, and N. Gordon, Eds. NewYork: Springer-Verlag, 2001: 197-223. [9] Storvik G. Particle filters in state space models with thepresence of unknown static parameters[J].IEEE Trans. onSignal Processing.2002, 50(2):281-289 [10] Wan E A and Nelson A T. Dual extended Kalman filtermethods. in Kalman Filtering and Neural Networks, S.Haykin, Eds. New York: John Wiley and Sons, Inc., 2001:123-173. [11] Arulampalam M S and Maskell S, et al.. A tutorial on particlefilters for online nonlinear/non-Gaussian Bayesian tracking[J].IEEE Trans. on Signal Processing.2002, 50(2):174-188 [12] Minvielie P and Marrs A D, et al.. Joint target tracking andidentification: part I: sequential Monte Carlo model-basedapproaches. 8th International Conference on InformationFusion. Philadelphia, USA: FUSION'2005: 25-29. [13] Ristic B and Farina A, et al.. Performance bounds andcomparison of nonlinear filters for tracking a ballistic objecton re-entry[J].IEE Proceedings on Radar, Sonar andNavigation.2003, 150(2):65-70 [14] 帕普里斯A, 佩莱S. 保铮等译. 概率、随机变量与随机过程.西安：西安交通大学出版社，2004: 70-72. [15] Kay S M. 罗鹏飞，张文明等译. 统计信号处理基础估计与检测理论. 北京：电子工业出版社，2006: 85-102. [16] Anderson B and Moore J. Optimal Filtering. EnglewoodCliffs, NJ: Prentice-Hall. 1979: 193-222. [17] Kitagawa G. A nonlinear smoothing method for time seriesanalysis. Statistica Sinica, 1991, 1(2): 371-388. [18] Chen E J. Simulation-based estimation of quantiles.Proceedings of the 31st conference on Winter simulation,Arizona, United States, 1999: 428-434. [19] Athans R and Berolini A. Suboptimal state estimation forcontinuous-time nonlinear systems from discrete noisymeasurements[J].IEEE Trans. on Automatic Control.1968,13(5):504-514 [20] Diaz-Garcia J A and Jaimez R G. Noncentral matrix variatebeta distribution. available from http: // www.cimat.mx/reportes/enlinea/I-06-06.pdf. 2006.12. 24. [21] Wagle B. Multivariate beta distribution and a test formultivariate normality. Journal of the Royal StatisticalSociety, 1968, Series B, 30(3): 511-516.

施引文献

资源附件(0)

访问统计