AN APPLICATION OF ARTIFICIAL NEURAL NETWORK TO COMPUTE THE RESONANT FREQUENCY OF E-SHAPED COMPACT MICROSTRIP ANTENNAS

Ali Akdagli – Abdurrahim Toktas – Ahmet Kayabasi – Ibrahim Develi

   An application of artificial neural network (ANN) based on multilayer perceptrons (MLP) to compute the resonant frequency of E-shaped compact microstrip antennas (ECMAs) is presented in this paper. The resonant frequencies of 144 ECMAs with different dimensions and electrical parameters were firstly determined by using SIE3D^{TM<\sup> software based on the method of moments (MoM), then the ANN model for computing the resonant frequency was built by considering the simulation data. The parameters and respective resonant frequency values of 130 simulated ECMAs were employed for training and the remaining 14 ECMAs were used for testing the model. The computed resonant frequencies for training and testing by ANN were obtained with the average percentage errors (APE) of .257 % and 0.523 %, respectively. The validity and accuracy of the present approach was verified on the measurement results of an ECMA fabricated in this study. Furthermore, the effects over the resonant frequency of the slots loading method onto the patch of ECMAs were investigated to explain the relationship between for miniaturization.}

Keywords: offline modeling, online modeling, RBF neural network, flux-linkage, switched reluctance motor (SRM)

[full-paper]