Prototipaje rápido de redes neuronales mediante el xPC Target de MATLAB
- Angel Javier Mazón Sainz-Maza
- Inmaculada Zamora Belver
- Koldobika Joseba Sagastabeitia Buruaga
- José Ángel Ezquerra Ventosa
- Víctor Valverde Santiago
- R. Vicente
Publisher: Asociación Española de Ingeniería de Proyectos (AEIPRO)
ISBN: 84-95809-22-2
Year of publication: 2005
Congress: CIDIP. Congreso Internacional de Ingeniería de Proyectos (8. 2004. Bilbao)
Type: Conference paper
Abstract
In the field of electrical engineering, they are many areas in which the artificial neural networks are being applied. An artificial neuronal network can be simulated by means of software executed in conventional computers. This constitutes the first step in the implementation. Nevertheless, a hardware implementation will have to be done if we want to take advantage of their capacity of massively parallel calculation, allowing their application in real time. A first step in solving this problem consists on carrying out the implementation by means of the use of "rapid prototyping techniques". Thus, the sotware Matlab offers a tool, the xPC-Target, that allows to simulate systems in real time and is able to develop the process of the solution´s prototyping. With this tool, the prototype of two types of artificial neural networks has been developed: the MLP and the LVQ networks. The obtained prototypes have been verified satisfactorily in the process of fault classification and fault localization in electrical transmission lines.