Title :
A neural-network-based system identification technique
Author :
Stubberud, A. ; Wabgaonkar, H. ; Stubberud, S.
Author_Institution :
California Univ., Irvine, CA, USA
Abstract :
A dynamic system identification technique based on neural networks is presented. The key idea is that the static training techniques can be applied to the dynamic case as well, provided the system is represented by a nonlinear state-space model. The feedforward-type neural network was trained using an extended Kalman filter to capture the input-output characteristics of a dynamic system
Keywords :
Kalman filters; feedforward neural nets; filtering and prediction theory; identification; learning (artificial intelligence); nonlinear systems; state-space methods; dynamic system identification; extended Kalman filter; feedforward-type neural network; input-output characteristics; neural-network-based system identification technique; nonlinear state-space model; static training techniques; Approximation algorithms; Artificial neural networks; Difference equations; Feedforward neural networks; Neural networks; Neurons; Nonlinear dynamical systems; Nonlinear equations; State estimation; System identification; Vectors;
Conference_Titel :
Decision and Control, 1991., Proceedings of the 30th IEEE Conference on
Conference_Location :
Brighton
Print_ISBN :
0-7803-0450-0
DOI :
10.1109/CDC.1991.261441
