EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Two-Stage Framework for Energy-Efficient Group Activation in RIS-Assisted Communication Systems

Hye-Yoon Jeong, Ki-Chang Tong, Dong-Woo Seo, Myung-Sun Baek, Gyeong-Ho Lee, Hyoung-Kyu Song

IEEE Access · 2026 · 14 · 108693-108713

DOI: 10.1109/access.2026.3713901 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Hye-Yoon Jeong und 5 weiteren Autor:innen. Die Arbeit nennt 31 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Research Foundation of Kore, Ministry of Education, Institute of Information & communications Technology Planning & Evaluation.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper investigates energy-efficient and reliable group-wise activation control for a reconfigurable intelligent surface (RIS)-assisted multi-user multiple-input single-output (MU-MISO) downlink under slot-level traffic dynamics. While activating all RIS groups can maximize the available reflecting aperture, it may waste non-negligible circuit and control power when traffic demand is intermittent or when multiple activation masks provide similar delivered goodput. Conversely, exhaustive per-slot evaluation of all non-empty group masks incurs prohibitive computational cost as the number of RIS groups increases. To address this tradeoff, a learning-assisted two-stage RIS group activation framework is developed. In the first stage, a graph neural network (GNN) exploits user-group interactions to screen and rank candidate activation masks, returning a compact top- $k$ set. In the second stage, each selected candidate is refined through a short physics-based iterative update to recover phase-optimization gain, followed by an energy efficiency (EE) gating rule with reliability-aware tie-breaking to avoid brittle low-power decisions. Simulation results with an $8\times 8$ RIS partitioned into four groups show that the proposed Graph-Assisted Two-Stage Energy-Gated RIS control framework, termed GATE-RIS, achieves 60.25 Mbit/J at $P_{\mathrm {tx}}=18$ dBm. This outperforms the always-on strategy at 50.47 Mbit/J and the rule-based baseline at 44.63 Mbit/J, while remaining close to the exhaustive EE oracle at 69.72 Mbit/J. In addition, GATE-RIS reduces the online runtime with an average $4.1-4.2\times $ speedup over exhaustive search. These results indicate that near-oracle EE can be achieved with bounded online complexity in the evaluated setting by combining learned candidate screening with lightweight physics-based refinement.

Förderung

  • National Research Foundation of Kore
  • Ministry of Education
  • Institute of Information & communications Technology Planning & Evaluation
  • Korea governmen
  • MSIT(Ministry of Science and ICT), Korea
  • IIT

Zitieren

APA

Jeong, H., Tong, K., Seo, D., Baek, M., Lee, G., & Song, H. (2026). A Two-Stage Framework for Energy-Efficient Group Activation in RIS-Assisted Communication Systems. IEEE Access, 14, 108693-108713. https://doi.org/10.1109/access.2026.3713901

Vancouver

Jeong H, Tong K, Seo D, Baek M, Lee G, Song H. A Two-Stage Framework for Energy-Efficient Group Activation in RIS-Assisted Communication Systems. IEEE Access. 2026;14:108693-108713. doi:10.1109/access.2026.3713901

BibTeX
@article{Jeong2026,
  author = {Jeong, Hye-Yoon and Tong, Ki-Chang and Seo, Dong-Woo and Baek, Myung-Sun and Lee, Gyeong-Ho and Song, Hyoung-Kyu},
  title = {A Two-Stage Framework for Energy-Efficient Group Activation in RIS-Assisted Communication Systems},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {108693--108713},
  doi = {10.1109/access.2026.3713901}
}
Kostenlos starten →