AI-Assisted Random Forest Recurrent Regression for Optimal Polarization Mode Selection in Hybrid Polarized MIMO Channels Using Shared Dual-Feed Antenna with Orthogonal Ports

Engineering Research · Springer Nature · 2026-11-13 · DOI: 10.1007/s13369-026-11795-w

Аннотация

In this study, we propose an AI-assisted approach for the optimal selection of polarization mode combinations in hybrid-polarized (HP) multiple-input multiple-output (MIMO) channels under varying noise conditions. Current research lacks unified frameworks for intelligent polarization selection and adaptive HP communication. To address this gap, an AI-assisted random forest recurrent regression (RFRR) model predicts the optimal polarization mode using channel metrics, improving signal quality and communication reliability. To realize the proposed HP communication framework, a novel shared dual-feed antenna with orthogonal ports (SDFAOP) MIMO is developed, enabling hybrid linear polarization (LP) and circular polarization (CP). The proposed SDFAOP employs two independent feeds per radiating element, supporting multiple polarization modes and multi-port operation. The proposed SDFAOP features a shared dual-feed structure with orthogonal ports, enabling the same radiating element to support both LP and CP for enhanced operational flexibility. Performance evaluation under diverse channel noise situations demonstrates the effectiveness of the proposed system. The RFRR-based optimal polarization mode selection is further analyzed using explainable AI (XAI) techniques to interpret the model’s decision-making process. The proposed SDFAOP MIMO antenna operating at 3.5 GHz is assessed using metrics such as $$S$$-parameters, axial ratio (AR), radiation patterns, total active reflection coefficient (TARC), envelope correlation coefficient (ECC), diversity gain (DG), channel capacity loss (CCL), mean effective gain (MEG), and efficiency. The results demonstrate excellent ECC (< 0.1), high DG (> 9.75 dB), low TARC, minimal CCL, and reduced cross-polarization. These findings highlight the superior accuracy of the AI-assisted mode selection, contributing to enhanced spectral efficiency and robustness in next-generation MIMO.

Читать у издателя →

Похожие статьи