Investigation of Holographic Beamforming via Dynamic Metasurface Antennas in QoS Guaranteed Power Efficient Networks

TL;DR

The paper addresses power-efficient downlink beamforming for a DMA-aided multi-user MISO system under SINR guarantees. It develops a holographic beamforming framework using semidefinite relaxation (SDR) and an alternating optimization between digital precoders $\{\mathbf{w}_m\}$ and DMA weights $\mathbf{Q}$, incorporating Lorentzian metasurface constraints and a near-field spherical-wave channel model. Key contributions include a novel SDR-based method that minimizes total transmit power under SINR constraints, a detailed analysis of Lorentzian constraint and reduced DoF effects through Monte Carlo simulations, and a comparative study against fully digital (FD) architectures across varying numbers of users and aperture configurations. The results provide practical design insights for DMA-enabled QoS-guaranteed networks, indicating how increasing DoF (e.g., more elements or denser spacing) mitigates the performance gap to FD in dense-user scenarios.

Abstract

This work focuses on designing a power-efficient network for Dynamic Metasurface Antennas (DMA)-aided multi-user multiple-input single-output (MISO) antenna systems. Power efficiency is achieved through holographic beamforming in a DMA-aided network, minimizing total transmission power while ensuring a guaranteed signal-to-noise-and-interference ratio (SINR) for multiple users in downlink. Unlike conventional MISO systems, which have well-explored beamforming solutions, DMA require specialized methods due to their unique physical constraints and wave-domain precoding capabilities. To achieve this, optimization algorithms relying on alternating optimization and semi-definite programming, are developed, including spherical-wave channel modelling of near-field communication. In this setup, the beamforming performance of DMA-aided precoding is analyzed in comparison to its optimal limits and traditional fully digital (FD) architectures, considering the effects of the Lorentzian constraints of metasurfaces and the degree of freedom (DoF) limitations due to a reduced number of RF chains. We demonstrate that the performance gap caused by DoF constraints becomes more significant as the number of users increases, highlighting the trade-offs of DMA in high-density wireless networks.

Figures (5)

• Figure 1: System configurations for different architectures: (a) FD, (b) DMA.
• Figure 2: Mean transmitted power versus minimum rate requirement at $K=2$.
• Figure 3: Mean transmitted power versus number of users ($K$) at $R_{\text{min},k}=$20 bps/Hz for $\forall k$.
• Figure 4: Comparison of different aperture weight distributions for optimizations OP1-FD, OP2-UW, and OP2-DMA $d_x=d_y=\lambda/2$; (a) ($K=1$), (b) ($K=2$).
• Figure 5: Mean transmitted power versus antenna spacing ($d_x$), and number of elements ($N$) at $K = 1$, $K = 2$, and $K = 4$.