A novel dynamic neural network structure for nonlinear system identification
| Authors: | Deng Jiamei, The University of Reading, United Kingdom
Becerra Victor, The University of Reading, United Kingdom
Nasuto Slawomir, The University of Reading, United Kingdom |
|---|
| Topic: | 3.2 Cognition and Control ( AI, Fuzzy, Neuro, Evolut.Comp.) |
|---|
| Session: | Neural Networks in Modelling and Control |
|---|
| Keywords: | recurrent neural networks, system identification, nonlinear systems |
|---|
Abstract
Dynamic neural networks are often used for nonlinear systemidentification. This paper presents a novel series-paralleldynamic neural network structure which is suitable for nonlinearsystem identification. A theoretical proof is given showing thatthis type of dynamic neural network is able to approximate finitetrajectories of nonlinear dynamical systems. Also, this neuralnetwork is trained to identify a practical nonlinear 3D cranesystem.
