Universidad, Ciencia y Tecnología
versión impresa ISSN 1316-4821
Resumen
ZAMBRANO, Alejandro; COLLAZO, Víctor; TRONCONE, Numan y RODRIGUEZ, Jesús. NIS: A Computer tool for closed-loop dynamical system identification through artificial neural networks. uct [online]. 2012, vol.16, n.64, pp.190-202. ISSN 1316-4821.
This paper shows the design and implementation of a computational tool for dynamical system identification by applying the technology of Artificial Neural Networks. The Systems Neuro-Identifier (NIS), has two main components: a stimulation and data acquisition hardware, and a human-machine interface. The NIS automatically generates first order transfer functions plus dead time (F.O.P.D.T.) and second order transfer functions if the system response is sub-damped. The tool has been applied in the identification of models under simulation environment and electrical circuits of first and second order. The models obtained matched the hypotheses regarding the dynamics of the plants identified, obtaining a cuadratic multivariable correlation factor of 0.9670 in the first order identified systems.
Palabras clave : Dynamical System Identification; Artificial Neural Networks; Human-Machine Interface.