Multi-Pair IRS-Aided Full-Duplex Relay Networks: Rate Analysis and Beamforming Design
Abstract
The intelligent reflecting surface (IRS) is a promising sixth-generation (6G) communication technology that can improve the wireless communication capacity in a cost-effective and energy-efficient manner by adjusting its passive reflectors to change the signal propagation. In this study, we investigate the multi-pair IRS-aided full-duplex relay (IFDR) network, wherein the relay is equipped with massive antennas. First, we determine the channel estimates of the cascaded channels and corresponding channel statistics for the IFDR system using the linear minimum mean square error (LMMSE) method. Second, we derive a closed-form lower bound for the achievable rate via maximum-ratio combining/transmission (MRC/MRT) processing and the use-and-then-forget (UatF) technique. This expression helps analyze the power scaling laws to provide better insight into the performance gain achieved by the IFDR system. Third, we propose an optimization algorithm to control the IRS phase-shift matrix and maximize the sum spectral efficiency using statistical channel knowledge. In particular, the problem is formulated as a fractional matrix programming problem. Subsequently, it is iteratively solved by using the majorization-minimization (MM) and square penalty methods. This approach can significantly reduce both the training and IRS configuration overhead. Lastly, through experiments, the numerical results show that the proposed control achieves better performance than the benchmark schemes.