Energy-Efficient Active Element Selection in RIS-aided Massive MIMO Systems

TL;DR

This work addresses EE optimization in active RIS-aided M-MIMO systems, formulating a unified framework to determine the number of active RIS elements needed to outperform fully passive RIS. The authors develop a low-complexity -Based amplitude/phase beamforming design that leverages Lagrangian dual transform and fractional programming to handle sum-of-ratios and non-convex constraints, with a semidefinite relaxation to recover a rank-one solution for the RIS coefficient vector $\boldsymbol{v}$. Key findings show up to a $120\%$ EE improvement over passive RIS and that active RIS can operate with roughly half the elements of a passive RIS while achieving comparable EE, given appropriate amplification power and element count (e.g., $N>49$ for full amplification utilization). The results underscore the practical viability of active RIS to mitigate double-fading effects, reduce channel estimation overhead, and enable compact RIS deployments, with future work extending to near-field regimes and multi-active-RIS configurations. Overall, the paper provides a rigorous optimization pipeline and actionable insights for deploying energy-efficient active RIS in future wireless networks.

Abstract

This chapter delves into the critical aspects of optimizing energy efficiency (EE) in active reconfigurable intelligent surface (RIS)-assisted massive MIMO (M-MIMO) wireless communication systems. We develop a comprehensive and unified theoretical framework to analyze the boundaries of EE within M-MIMO systems integrated with active RIS while adhering to practical constraints. Our research focuses on a formulated EE optimization problem aiming to maximize the EE for active RIS-assisted M-MIMO communication systems. Our goal is to strategically find the number of active RIS elements for outperforming the EE attainable by an entirely passive RIS. Besides, the proposed novel solution has been tailored to the innovative problem. The formulation and solution design consider analytical optimization techniques, such as lagrangian dual transform (LDT) and fractional programming (FP) optimization, facilitating the effective implementation of RIS-aided M-MIMO applications in real-world settings. In particular, our results show that the proposed algorithm can provide up to 120% higher EE than the entirely passive RIS. Besides, we found that the active RIS can operate with less than half of the reflecting elements for the entirely passive RIS. Finally, in view of active RIS achieving the complete utilization of amplification power available, it should be equipped with a reasonable number of reflecting elements above N = 49.

Figures (5)

• Figure 1: Average vs the transmit power budget [dBm] at the ($P_{\rm{TX}}$). Performance evaluation of the proposed algorithm for the active vs entirely passive and random phase shift/precoding, with $N=25$.
• Figure 2: Percentage of power amplification utilized vs the number of reflecting elements for the proposed algorithm. In this setup, we set $P_{\rm TX} = 35$ dBm.
• Figure 3: CDF of the average value of amplitude obtained with the proposed algorithm. Here we set $P_{\rm TX} = 35$ dBm.
• Figure 4: Average vs number of RIS reflecting units. Performance evaluation of the proposed algorithm for the active vs entirely passive with fixed number of elements $N=64$, for $P_{\rm TX} = 35$ dBm and two different number of , $K=5$ and $K=10$.
• Figure 5: Average vs number of iterations for the proposed algorithm. Performance evaluation of the proposed algorithm for the active with fixed number of elements $N=25$, for $P_{\rm TX} = \{10, 35, 50\}$ dBm and for fixed transmit power budget $P_{\rm TX} = 35$ dBm and $N = \{49, 64\}$.