Título: Identificação de Processo Real Não Linear do Tipo Hammerstein utilizando Redes Neurais com Dinâmica Externa e com Dinâmica Interna
Autores: Oliveira, R. C. L. de; Soares, R. P. O.; Costa, C. E. U.
Resumo: Neural Networks have been used to identify nonlinear dynamic systems because they are genuine nonlinear black box models and they have ability to approximate complicated nonlinear relations. This paper presents two neural net models, one with external dynamic, represented by a regression vector and the other with internal dynamic represented by dynamic neurons which are used to identify nonlinear systems. They has a feedforward topology and the learning algorithm used is backpropagation. Results attained in the identification of a real block-oriented nonlinear dynamic process are used to evaluate characteristic of each these neural net models.
Código DOI: 10.21528/CBRN2001-113
Artigo em pdf: 5cbrn_113.pdf
Arquivo BibTex: 5cbrn_113.bib