Nonlinear Identification Based on Diagonal Recurrent Neural Network and Particle Filter

Author

Xiaolong, Deng ; Pingfang, Zhou

Author_Institution

Dept. of Mech. Eng., Jiangsu Coll. of Inf. Technol., Wuxi, China

Volume

1

fYear

2009

fDate

14-16 Aug. 2009

Firstpage

217

Lastpage

221

Abstract

Diagonal recurrent neural network (DRNN) is widely applied to nonlinear identification. In this paper, the extended Kalman filter and particle filter are firstly combined to train DRNN. Utilizing time windows, a method to evaluate the dynamical performance of DRNN is presented. Network weights of particles are optimized by the resampling algorithm. The high convergent speed and high training precision are obtained by the new algorithm. Simulation results of the nonlinear dynamical identification verify the validity of the new algorithm.

Keywords

Kalman filters; algorithm theory; nonlinear dynamical systems; particle filtering (numerical methods); recurrent neural nets; diagonal recurrent neural network; extended Kalman filter; high convergent speed; high training precision; new algorithm validity; nonlinear dynamical identification; nonlinear identification based; particle filter; resampling algorithm; Artificial neural networks; Chaos; Computer networks; Delay estimation; Educational institutions; Mechanical engineering; Neurofeedback; Neurons; Particle filters; Recurrent neural networks; diagonal recurrent neural network; nonlinear identification; particle filter; the extended Kalman filter; training algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Natural Computation, 2009. ICNC '09. Fifth International Conference on

Conference_Location

Tianjin

Print_ISBN

978-0-7695-3736-8

Type

conf

DOI

10.1109/ICNC.2009.496

Filename

5367167