|
NONLINEAR IDENTIFICATION VIA VARIABLE STRUCTURE RECURRENT NEURAL NETWORKS
Edgar N. Sanchez and Ramon A. Felix*
* CINVESTAV, Unidad Guadalajara, Apartado Postal 31-438,
Plaza La Luna, Guadalajara, Jalisco, C.P. 45091, Mexico,
e-mail:sanchez@gdl.cinvestav.mx
In this paper, we present a new approach for nonlinear identification using Variable Structure Recurrent Neural Networks (VSRNN). We propose a neural network identifier, whose structure changes depending on the error. In this way, a trade off between identification error and computational complexity is achieved.
Keywords: Dynamic Neural Networks, Variable Stucture Systems, Nonlinear systems, Identification, Lyapunov methodology
Session slot T-Th-E04: Neural and fuzzy Identification/Area code 3e : Fuzzy and Neural Systems
 | 