Applications of neural networks in learning of dynamical systems
| Autor: | S.R. Chu, R. Shoureshi |
|---|---|
| Rok vydání: | 1992 |
| Předmět: |
Dynamical systems theory
Artificial neural network Nonlinear system identification Computer science business.industry General Engineering System identification Machine learning computer.software_genre Dynamical system Hopfield network Identification (information) Cellular neural network Artificial intelligence Types of artificial neural networks business computer |
| Zdroj: | IEEE Transactions on Systems, Man, and Cybernetics. 22:161-164 |
| ISSN: | 0018-9472 |
| DOI: | 10.1109/21.141320 |
| Popis: | One of the immediate applications of neural networks in the engineering field is pattern recognition and its extension to system identification. Three unique features of neural networks, namely, learning, high-speed processing of massive data, and the ability to handle signals with degrees of uncertainty, make such networks attractive to dynamical systems. The first step in analyzing such systems is to learn the dynamics of the system, i.e., system identification. A time-domain approach using a Hopfield network and a frequency-domain approach using spectral decomposition for identification of dynamical systems are presented. Simulation results are discussed. > |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|