SIM-Enabled Hybrid Digital-Wave Beamforming for Fronthaul-Constrained Cell-Free Massive MIMO Systems

Eunhyuk Park, Seok-Hwan Park, Osvaldo Simeone, Marco Di Renzo, Shlomo ShamaiIEEE Trans. Wireless Communications 2026 · DOI

TL;DR

This paper introduces a novel hybrid beamforming framework for cell-free massive MIMO (CF-mMIMO) systems that integrates intelligent metasurfaces (SIMs) with conventional digital processing. This approach significantly reduces RF chain requirements and fronthaul demands, enabling more cost-effective and scalable CF-mMIMO deployments while achieving performance close to fully-digital systems.

What Problem Does This Paper Address?

Cell-free massive MIMO (CF-mMIMO) systems promise enhanced wireless connectivity by distributing numerous access points (APs) across a service area. However, equipping each AP with large-scale antenna arrays (LAAs) to improve coverage leads to substantial implementation costs due to the need for numerous RF chains. Furthermore, the high data rates required for signal processing and coordination between APs and a central processor (CP) create significant fronthaul capacity demands, which are expected to grow with future network generations. This paper tackles the challenge of achieving high performance in CF-mMIMO systems while mitigating these cost and capacity limitations.

Methods

The core innovation is a hybrid digital-wave beamforming framework that leverages Stacked Intelligent Metasurfaces (SIMs). SIMs, composed of programmable metasurfaces, can manipulate electromagnetic waves to perform beamforming directly in the wave domain. This allows APs to achieve significant beamforming gains with fewer RF chains. The paper proposes a joint optimization framework that simultaneously optimizes:

  • Digital beamforming: Standard signal processing at the CP and APs.
  • Wave-domain beamforming: Phase shifts controlled by the SIMs at the APs.
  • Fronthaul compression: Efficiently encoding signals for transmission over limited fronthaul links.

To handle the resulting non-convex optimization problems, the authors develop Alternating Optimization (AO)-based algorithms. These algorithms iteratively optimize different sets of variables (digital processing, wave-domain beamforming, and fronthaul compression) until convergence, ensuring a monotonic increase in the system's weighted sum-rate. For specific subproblems, they employ techniques like matrix Lagrangian duality and Fenchel's inequality to transform non-convex problems into convex ones or to derive closed-form solutions.

Datasets

Not discussed in the paper. The paper focuses on theoretical formulation and simulation-based evaluation.

Key Insights and Ideas

  • SIMs as a Game-Changer for RF Chain Reduction: The wave-domain beamforming capability of SIMs is shown to be highly effective in reducing the number of RF chains needed per AP, directly addressing a major cost driver in CF-mMIMO.
  • Synergy Between Digital and Wave-Domain Processing: The paper demonstrates that jointly optimizing both digital and wave-domain beamforming yields significant performance gains compared to optimizing them independently or relying on random wave-domain configurations.
  • Fronthaul Compression is Crucial: The optimization of fronthaul compression is shown to be critical, especially when combined with SIM-enabled wave-domain beamforming, leading to substantial performance improvements.
  • Deep SIMs Offer Near Fully-Digital Performance: With a sufficient number of metasurface layers (deeper SIMs), the proposed hybrid scheme can approach the performance of a fully-digital system, even with significantly fewer RF chains.
  • Cost-Effectiveness: The proposed approach offers a pathway to more cost-effective CF-mMIMO deployments by reducing hardware complexity (RF chains) and computational load.

Key Results

  • The proposed hybrid digital-wave beamforming scheme significantly outperforms conventional schemes that rely on randomly set wave-domain beamformers or restrict digital beamforming to simple power control.
  • In the high SNR regime, the proposed schemes with sufficiently deep SIMs achieve performance close to fully-digital beamforming, using substantially fewer RF chains.
  • The complexity analysis shows that the proposed algorithms have lower computational complexity compared to fully-digital schemes, despite the added optimization of wave-domain variables.
  • Numerical results demonstrate substantial performance gains from optimizing fronthaul compression jointly with SIM-enabled wave-domain beamforming.

Limitations and Future Work

  • Perfect CSI Assumption: The current work assumes perfect channel state information (CSI). Future work could explore robust designs under imperfect CSI.
  • Analog Beamforming: The paper focuses on digital and wave-domain beamforming, leaving the joint optimization involving analog beamforming for future research.
  • Future Research Directions: The authors plan to extend their work to:
  • Hybrid digital-wave channel estimators.
  • Robust hybrid beamforming designs under imperfect CSI.
  • Integrating reconfigurable antenna techniques.
  • Extending uplink-downlink duality results to SIM-based architectures.
  • Optimizing inter-layer transmission matrices using flexible intelligent metasurfaces.
  • Developing low-complexity designs using advanced efficient algorithms.

Summary updated

BibTeX
@article{park2026sim,
  author = {Eunhyuk Park and Seok-Hwan Park and Osvaldo Simeone and Marco Di Renzo and Shlomo Shamai},
  title = {{SIM-Enabled Hybrid Digital-Wave Beamforming for Fronthaul-Constrained Cell-Free Massive MIMO Systems}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2026},
  doi = {10.1109/twc.2025.3628306},
}
Figure