Effective modeling of metasurface‐loaded circularly polarized dielectric resonator‐based MIMO antenna for sub‐6.0‐GHz band using machine learning algorithms

Abstract

SummaryIn this article, a dual‐port dielectric resonator antenna (DRA) is modeled using machine learning (ML) algorithms, that is, deep neural network (DNN), random forest, and XGBoost. The unique properties of the proposed article are as follows: (i) Two different diversity techniques, that is, pattern (with the help of metallic wall) and polarization (mirror image of the aperture), improve the isolation value between the ports; (ii) ML algorithms are used to optimize and predict the reflection coefficient as well as mutual coupling of the proposed antenna. The accuracy of ML algorithms is verified by using the HFSS EM simulator and experimental validation. Error is less than 1%–2% between the value predicted from ML algorithms and HFSS/experimental results. The proposed design is working well in between 2.4 and 4.02 GHz with a 3‐dB axial ratio from 2.84 to 2.95 GHz. All these features make the radiator employable to the sub‐6.0‐GHz frequency band.