Machine learning‐enabled two‐port wideband MIMO hybrid rectangular dielectric resonator antenna for n261 5G NR millimeter wave

Summary In this article, a two‐port multiple‐input multiple‐output (MIMO) hybrid rectangular dielectric resonator antenna (DRA) with machine learning (ML) approach for the n261 5G New Radio (NR) application is presented. The proposed antenna is designed on an RT/duroid 5880 (Ɛr = 2.2) substrate activated by 50 Ω, L‐shaped microstrip slot feeds beneath both DRAs. The isolation is more than 19 dB, and the gain is 10 dBi in the operating frequency range. The proposed antenna is optimized through knowledge‐based neural networks (KBNN), artificial neural networks (ANNs), and ML. The optimal design parameters of the proposed antenna are accomplished using the ML optimization approach, which includes ridge regression, ANNs, and KBNN. KBNN ML techniques provide 96.88% accuracy and correctly predict the S‐parameters of the proposed antenna. The MIMO diversity parameters like envelope correlation coefficient (ECC), diversity gain (DG), total active reflection coefficient (TARC), and channel capacity loss (CCL) are calculated and found within the limits. Hence, the proposed antenna is used for 5G NR mm‐wave application.