Systems identification using recurrent asymptotically stable neural networks

Author

Jubien, Chris M. ; Dimopoulos, Nikitas J.

Author_Institution

Dept. of Electr. & Comput. Eng., Victoria Univ., BC, Canada

Volume

2

fYear

1993

fDate

19-21 May 1993

Firstpage

610

Abstract

A training procedure for a class of neural networks that are asymptotically stable is presented. The training procedure is a gradient method which adapts the interconnection weights as well as the relaxation constants and the slopes of the activation functions used so as to minimize the error between the expected and obtained responses. A method for assuring that stability is maintained throughout the training procedure is also given. Such a network was used to identify the dynamic behavior of several nonlinear dynamical systems, a PUMA 560 robot and a boat based on collected rudder/heading data

Keywords

asymptotic stability; identification; learning (artificial intelligence); nonlinear dynamical systems; recurrent neural nets; activation functions; boat; gradient method; identification; interconnection weights; nonlinear dynamical systems; recurrent asymptotically stable neural networks; relaxation constants; robot; training procedure; Cost function; Gradient methods; Neural networks; Neurons; Nonlinear dynamical systems; Nonlinear systems; Recurrent neural networks; Robots; Stability; System identification;

fLanguage

English

Publisher

ieee

Conference_Titel

Communications, Computers and Signal Processing, 1993., IEEE Pacific Rim Conference on

Conference_Location

Victoria, BC

Print_ISBN

0-7803-0971-5

Type

conf

DOI

10.1109/PACRIM.1993.407287

Filename

407287