Title: Software module 'pentagonal dipole' based on artificial neural networks

Authors: Marija Milijić; Zoran Stanković; Ivan Milovanović; Aleksandar Nešić

Addresses: Faculty of Electronic Engineering, University of Niš, Aleksandra Medvedeva 14, 18 000 Niš, Serbia. ' Faculty of Electronic Engineering, University of Niš, Aleksandra Medvedeva 14, 18 000 Niš, Serbia. ' Faculty of Informatics and Computing, University Singidunum, Belgrade, Serbia. ' IMTEL Komunikacije a.d., Bulevar Mihajla Pupina 165b, 11070 Novi Beograd, Serbia

Abstract: Software module 'pentagonal dipole' is design to calculate pentagonal dipole resonant frequency versus its dimension and substrate dielectric constant. The software is based on neural model that is trained and tested by results obtained by WIPL-D software package. Using Method of Moments, WIPL-D gives results of high accuracy, even though it can have a long simulation time. As alternative methods to WIPL-D software, artificial neural network can achieve both great simulation speed and the satisfactory accuracy. The great speed of neural model can give good base for introducing neural model in the optimisation process of planar structures or in applications which simulation has to be finished in certain period of time. Thanks to its ability of integration in user-friendly software, proposed neural model is used to develop software module 'pentagonal dipole' that features with great speed and accuracy as well as WIPL-D software.

Keywords: printed antenna; pentagonal dipole; artifical neural networks; ANNs; resonance frequency; modelling; simulation.

DOI: 10.1504/IJRIS.2012.046494

International Journal of Reasoning-based Intelligent Systems, 2012 Vol.4 No.1/2, pp.73 - 81

Published online: 10 Apr 2012 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article