EUVIMEDEUVIMED entdecken
Crossref

Two-Timescale Deep Optimization of Positioning and Beamforming in Movable Antenna Arrays

Fengyu Liang, Yunlong Cai, An Liu, Benoit Champagne

IEEE Transactions on Wireless Communications · 2026 · 25 · 14242-14258

DOI: 10.1109/twc.2026.3675048 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Wireless Communications, 2026. Verfasst von Fengyu Liang und 3 weiteren Autor:innen. Die Arbeit nennt 51 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Peng Cheng Laboratory Project, Zhejiang Provincial Key Laboratory of Multi-Modal Communication Networks and Intelligent Information Processing.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In this study, we investigate a downlink multiuser multiple-input multiple-output (MU-MIMO) system employing a two-dimensional (2D) movable antenna (MA) array. We propose a two-timescale optimization framework to jointly optimize antenna position vector (APV) and beamforming for sum-rate maximization, addressing hardware limitations that restrict real-time antenna adjustments. Specifically, APV is updated based on long-term channel statistics at the start of each coherence block, while beamforming is optimized per time slot using short-term information. An efficient stochastic successive convex approximation (SSCA)-based algorithm is developed for joint optimization. To enhance performance and reduce complexity, we propose a deep-unfolding neural network (DUNN) integrating non-linear activations and first-order Taylor approximations for matrix inversions. Furthermore, we introduce a meta-learning approach for improved initialization and rapid convergence, along with an online adaptation mechanism for continuous adjustment to changing channel conditions. Simulation results show that our proposed approach improves sum-rate compared to conventional MIMO systems with uniform arrays, and the DUNN outperforms the SSCA-based algorithm. Meanwhile, meta-learning and online adaptation framework enables network to adapt to channel changes better and converge faster.

Förderung

  • National Natural Science Foundation of China
  • Peng Cheng Laboratory Project
  • Zhejiang Provincial Key Laboratory of Multi-Modal Communication Networks and Intelligent Information Processing

Zitieren

APA

Liang, F., Cai, Y., Liu, A., & Champagne, B. (2026). Two-Timescale Deep Optimization of Positioning and Beamforming in Movable Antenna Arrays. IEEE Transactions on Wireless Communications, 25, 14242-14258. https://doi.org/10.1109/twc.2026.3675048

Vancouver

Liang F, Cai Y, Liu A, Champagne B. Two-Timescale Deep Optimization of Positioning and Beamforming in Movable Antenna Arrays. IEEE Transactions on Wireless Communications. 2026;25:14242-14258. doi:10.1109/twc.2026.3675048

BibTeX
@article{Liang2026,
  author = {Liang, Fengyu and Cai, Yunlong and Liu, An and Champagne, Benoit},
  title = {Two-Timescale Deep Optimization of Positioning and Beamforming in Movable Antenna Arrays},
  journal = {IEEE Transactions on Wireless Communications},
  year = {2026},
  volume = {25},
  pages = {14242--14258},
  doi = {10.1109/twc.2026.3675048}
}
Kostenlos starten →