基于无先导卡尔曼滤波的RBFN训练算法研究

张海涛;李大字*;靳其兵;耿延睿

北京化工大学学报(自然科学版) ›› 2007, Vol. 34 ›› Issue (2) : 221-224.

PDF(218 KB)
欢迎访问北京化工大学学报(自然科学版),今天是 2025年5月5日 星期一
Email Alert  RSS
PDF(218 KB)
北京化工大学学报(自然科学版) ›› 2007, Vol. 34 ›› Issue (2) : 221-224.
研究简报

基于无先导卡尔曼滤波的RBFN训练算法研究

  • 张海涛;李大字*;靳其兵;耿延睿
作者信息 +

Training radial basis neural networks with the unscented Kalman filter

  • ZHANG HaiTao;LI DaZi;JIN QiBing;GENG YanRui
Author information +
文章历史 +

摘要

提出了应用无先导卡尔曼滤波器(UKF)来训练径向基神经网络(RBFN)的新方法。与广义卡尔曼滤波器(EKF)和双重卡尔曼滤波器(DEKF)对函数的一阶近似不同,UKF对非线性函数采用二阶近似展开,而且最重要的一点是不必求取系统的雅克比矩阵,从而大大减小计算量。本文对时间序列预测及分类问题进行了仿真,结果证实了该方法的有效性和快速性。

Abstract

A new method is proposed for training radial basis function networks (RBFN) using the unscented Kalman filter (UKF). In contrast to the extended Kalman filter (EKF) and the dual extended Kalman filter (DEKF), which extend the nonlinear functions using a first order approximation, the UKF uses a second order approximation. The most important consequence is that the algorithm does not require the Jacobi matrix of the system to be calculated, thus reducing the calculation complexity and resulting in considerable savings in time. Simulation results in the fields of chaotic time series prediction and classification problems demonstrate both the validity and faster speed of the proposed method.

引用本文

导出引用
张海涛;李大字*;靳其兵;耿延睿. 基于无先导卡尔曼滤波的RBFN训练算法研究[J]. 北京化工大学学报(自然科学版), 2007, 34(2): 221-224
ZHANG HaiTao;LI DaZi;JIN QiBing;GENG YanRui. Training radial basis neural networks with the unscented Kalman filter [J]. Journal of Beijing University of Chemical Technology, 2007, 34(2): 221-224

参考文献

[1]SCHWENKER F, KESTLER H A, PALM G, et al. Three learning phase for radial basis funciton networks[J]. Neural Networks, 2001, 14(4-5): 439-458.
[2]SARIMVEIS H, DOGANIS P, ALEXANDRIDIS A, et al. A classification technique based on radial basis function neural networks[J]. Advances in Engineering Software, 2006,37(4): 218-221.
[3]KARAYIANNI N. Reformulated radial basis neural networks trained by gradient descent[J]. Neural Networks, 1999, 3(5):657-671.
[4]PIOVOSO M, PHILLIP A, LAPLANTE L, et al. Kalman filter recipes for realtime image processing[J]. RealTime Imaging, 2003, 9(6):433-439.
[5]SIMON D. Kalman filtering for fuzzy discrete time dynamic systems[J]. Applied Soft Computing Journal, 2003, 3(3): 191-207.
[6]BOJE E, PETRICK M. Application of the extended Kalman filter to a lysine hydrochlorination process[J]. Control Engineering Practice, 2000, 8(3): 291-297.

[7]JULIER S, HLMANN J. A new approach for filtering nonlinear systems[C]. Piscataway, NJ, USA :IEEE Press, 1995: 1628-1632.
[8]LI Peihua, ZHANG Tianwen, MA Bo. Unscented kalman filter for visual curve tracking[J]. Image and Vision Computing, 2004, 22(2): 157-164.
[9]SIMON D. Training radial basis neural networks with the extended kalman filter[J]. Neurocomputing, 2002, 48(1-4): 455-475.
[10]IULIAN B C. RBF networks training using a dual extended [JP2]kalman filter[J]. Neurocomputing, 2002, 48(1-[JP]4): 609-622.
[11]ROMANENKO A, CASTRO JOSE A A M. The unscented filter as an alternative to the EKF for nonlinear state estimation: a simulation case study[J]. Computers and Chemical Engineering, 2004, 28(3): 347-355.

PDF(218 KB)

3260

Accesses

0

Citation

Detail

段落导航
相关文章

/