ABSTRACT This paper describes the design, fabrication, and experimental validation of a metasurface‐suspended two‐port dielectric resonator antenna (DRA) with integrated machine learning (ML)–based performance prediction for millimeter‐wave applications. The dielectric resonator's circular polarization is excited by an asymmetrical plus‐shaped aperture, and strong interport isolation is ensured by an orthogonal port arrangement. The antenna directivity and gain are increased by adding a double‐negative (DNG) metasurface superstrate. The suggested design concurrently achieves better gain (≈13 dBi), larger impedance bandwidth (1.65 GHz), and improved port isolation (> 30 dB) in comparison to previously published mm‐wave dielectric MIMO antennas, indicating a significant overall performance increase. The |S 11 | response of the antenna is successfully predicted using two ML approaches: Artificial Neural Network (ANN) and Random Forest (RF). The design is validated by measuring a manufactured prototype. The suggested antenna is a strong contender for n257/n258 mm‐wave 5G communication systems because it works in the 32.18–33.83 GHz range with steady circular polarization and excellent MIMO diversity features.
Shree et al. (2026) studied this question.