非线性非高斯结构系统识别的粒子滤波方法

英文篇名：Particle Filtering Method for Nonlinear Non-Gaussian Structural System Identification
中文刊名：江西科学
英文刊名：Jiangxi Science
作者：张伟 ; 李烨 ; 杨晓楠
英文作者：ZHANG Wei1 ; LI Ye2 ; YANG Xiao-nan1(1.Research Institute of Structural Engineering and Disaster Reduction ; Tongji University ; Shanghai 200092 PRC ; 2.East China Architectural Design & Research Institute Co. ; Ltd. ; Shanghai 200002 PRC)
中文关键词：结构损伤识别 ; 粒子滤波 ; 扩展卡尔曼滤波 ; 非线性 ; 非高斯
英文关键词：Structural system identification ; Particle filtering ; Extend kalman filtering ; Nonlinear ; Non-Gaussian
出版日期：2008-06-15
机构：同济大学结构工程与防灾研究所,华东建筑设计研究院有限公司,同济大学结构工程与防灾研究所上海200092,上海200002,上海200092
年：2008
期：03
出版单位：江西科学

摘要

采用粒子滤波方法(PF方法)在非高斯噪声条件下对非线性系统进行参数识别。传统扩展卡尔曼滤波(EKF)方法具有高斯噪声假设与非线性系统线性化的缺陷,PF方法可以克服EKF方法的缺点;因此在系统识别中具有很强的鲁棒性,更适合进行非线性结构系统参数识别。数值仿真结果发现PF方法的系统识别精度高于EKF方法,证明PF方法在非线性非高斯结构系统识别中的有效性。
A particle filtering(PF)method is utilized to identify a nonlinear structural system with non-Gaussian noise.Traditional extend Kalman filtering(EKF)method has some disadvantages in Gaussian noise hypothesis and to linearize nonlinear system,but the PF method can conquer such disadvantages of EKF method.Therefore,the PF method has great robust ability.It is suitable to nonlinear non-Gaussian structural parameter estimation.The numerical simulations confirm effectiveness of the proposed method for the online structural system identification.

引文

[1]冯新,李国强,周晶.土木工程结构健康诊断中的统计识别方法综述[J].地震工程与工程振动,2005,25(2):105-113.
    [2]Gordon N J,Salmond D J,Smith A F M.Novel ap-proach to nonlinear/non-Gaussian Bayesian state es-timation[J].IEE Proceedings-F,1993,140(2):107-121.
    [3]Arulampalam MS,Maskell S,Gordon N,et al.Atutori-al on particle filters for online nonlinear/non-Gaussi-an Bayesian tracking[J].IEEE Transactions on SignalProcessing,2002,50(2):174-192.
    [4]Doucet A,Godsill S,Andrieu C.On sequential MonteCarlo sampling methods for Bayesian filtering[J].Sta-tistics and Computing,2000,10(3):197-210.
    [5]雷明,韩崇昭,肖梅.扩展卡尔曼粒子滤波算法的一种修正方法[J].西安交通大学学报,2005,39(8):824-827.
    [6]Doucet J F G.de Freitas,Gordon N J.An introductionto sequential Monte Carlo methods[M].SequentialMonte Carlo Methods in Practice,New York:Springer-Verlag,2001.
    [7]Beadle E R,Djuric P M.A fast weighted Bayesianbootstrap filter for nonlinear model state estimation[J].IEEE Transactions on Aerospace and ElectronicSystems,1997,33(1):338-349.
    [8]Wen Y K.Method for random vibration of hystereticsystems[J].ASCE Journal of the Engineering Mechan-ics Division,1976,102(2):249-263.