Joint Optimization Design for Near-Field IRS-Assisted Modular XL-MIMO Systems

Existing research on intelligent reflective surface (IRS) primarily focuses on far-field communication scenarios. Even in studies involving large-scale IRS-assisted systems, the far-field communication assumption remains prevalent, potentially leading to inaccurate evaluations of communication performance. To this end, the paper introduces a modularized extremely large-scale IRS (XL-IRS) and investigates its performance in supporting near-field communications. Specifically, a comprehensive channel model is developed for the near-field communication scenario assisted by the modularized XL-IRS. To maximize the weighted sum rate of the system, an alternating optimization algorithm is proposed to jointly optimize the transmit precoding at the base station and phase shift matrices of the modularized XL-IRS. We subsequently introduce an alternating optimization algorithm that produces a closed-form solution. Simulation results provide strong evidence supporting the proposed alternating optimization algorithm achieves satisfactory coverage and rate performance. Additionally, it can be observed that as the distance between modules increases, the system's weighted achievable sum rate decreases and that near-field users achieve a higher weighted sum rate compared to far-field users.