Work

User-Centric Hardware Acceleration for Energy-Efficient Cell-Free Massive MIMO Systems

Year: 2026

Type: conference-paper

Abstract: The rapid increase in mobile data traffic necessitates a highly energy-efficient radio access network (RAN) architecture. In cell-free massive multiple-input multiple-output (CF-mMIMO) systems, hardware accelerators (ACCs) for physical-layer signal processing consume considerable static power because they must remain active to support control functions even during low-traffic periods. To address this inefficiency, we propose a user-centric hardware acceleration (UC-HA) architecture that separates signal processing into control-channel and per-user data-channel, and dynamically activates per-user ACCs according to user activity. A 5G outdoor testbed implementing UC-HA validates its feasibility and shows comparable throughput performance to conventional acceleration. Large-scale simulations in a 1 km2urban network calibrated by real power measurements demonstrate up to a 37% reduction in daily energy consumption compared with network-centric architectures. These results indicate that UC-HA constitutes a scalable and sustainable hardware platform for future green CF-mMIMO deployments.

Source: International Conference on Communications

DOI: 10.1109/icc59461.2026.11587649

Authors: Yu Tsukamoto, Takahide Murakami, Amr Amrallah, Akio Ikami

Institution: KDDI Research (Japan)

Language: English

Topic: Advanced MIMO Systems Optimization

Subfield: Electrical and Electronic Engineering

Field: Engineering

Domain: Physical Sciences

SDGs: Industry, innovation and infrastructure, Affordable and clean energy

Open Access status: hybrid

Funder: National Institute of Information and Communications Technology