Enhancing Indoor and Outdoor THz Communications with Beyond Diagonal-IRS: Optimization and Performance Analysis

TL;DR

This paper explores the use of Beyond Diagonal IRS to enhance THz communication systems by jointly optimizing beamforming and phase shifts, significantly improving data rates for indoor and outdoor users.

Contribution

It introduces a novel optimization framework for hybrid reflective/transmissive IRS in THz systems, employing an iterative algorithm to maximize system sum rate.

Findings

Achieves approximately 30.5% and 69.9% rate improvements over benchmark schemes.

Outperforms STAR-IRS system by 76.99%.

Highlights the critical role of IRS elements in system performance.

Abstract

This work investigates the application of Beyond Diagonal Intelligent Reflective Surface (BD-IRS) to enhance THz downlink communication systems, operating in a hybrid: reflective and transmissive mode, to simultaneously provide services to indoor and outdoor users. We propose an optimization framework that jointly optimizes the beamforming vectors and phase shifts in the hybrid reflective/transmissive mode, aiming to maximize the system sum rate. To tackle the challenges in solving the joint design problem, we employ the conjugate gradient method and propose an iterative algorithm that successively optimizes the hybrid beamforming vectors and the phase shifts. Through comprehensive numerical simulations, our findings demonstrate a significant improvement in rate when compared to existing benchmark schemes, including time- and frequency-divided approaches, by approximately $30.5\%$ and $69.9\%$ respectively and even outperforms the STAR-IRS system by $76.99\%$. This underscores the significant influence of IRS elements on system performance relative to that of base station antennas, highlighting their pivotal role in advancing the communication system efficacy.