Artificial Neural Network Established Hexagonal Ring- MPA Intended for Wideband Application

Abstract

Artificial neural networks (ANNs) have established eminent strategies for registering distinctive execution boundaries of microstrip patch antennas (MPAs). In this paper, a hexagonal ring (HR)- MPA is assessed for wideband by exploring ANN paradigm. The parametric inquests of diverse radiating configurations have also been performed to accomplish HRMPA. The proposed work illustrates an ANN paradigm configured with multilayer perceptron (MLP) for anticipating the efficient dimension of the radiator, ground plane, and slot for achieving the desired level of bandwidth (BW) and radiation efficiency (Reff) for wideband operation. For the analysis of the ANN paradigm, 303 examples are taken into consideration, whereas 213-45-45 example/datasets are derived for training, validation, and testing, respectively. The example sets are established by constructing the HRMPA by a high frequency structure simulator (HFSS). The simulated antenna structure provides 138.18% bandwidth above the frequency extent of 2.76–15.1 GHz with more than 80% radiation efficiency for the intact band. The outcomes attained utilizing ANN are compared by the simulation outcomes which are very close contract.