A neural network approach to a system identification problem is presented. Traditional system identification techniques require knowledge of the model structure before parameter estimation methods can be applied. This approach requires less a priori information and the knowledge of model structure is not essential. A simulation study on a power system has demonstrated the application of this technique
Published in:
Control Theory and Applications, IEE Proceedings -
(Volume:142
,
Issue:
4
)
Date of Publication: Jul 1995