EUVIMEDEUVIMED entdecken
Crossref

Constellation Design for Active-Set-Aware Digital Over-the-Air Computation

Zeyang Li, Carlo Fischione

IEEE Transactions on Communications · 2026 · 74 · 15703-15718

DOI: 10.1109/tcomm.2026.3745138 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Communications, 2026. Verfasst von Zeyang Li und 1 weiteren Autor:innen. Die Arbeit nennt 36 Literaturangaben. Gefördert von: Swedish Foundation for Strategic Research (SSF) through the Software Artificial Intelligence for Communications (SAICOM) project, Knut and Alice Wallenberg Foundation (KAW) through the Turning the Air into an AI Computer (TAIRCOMP) project, Digital Futures, and in part by Vinnova (Sweden’s Innovation Agency) through the SYNCRO6G project.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper studies active-set-aware digital over-the-air computation (AirComp), where the receiver is required to recover not only an aggregate numerical value but also the set of active transmitting devices from the same superimposed signal. Unlike existing AirComp designs that focus primarily on aggregation recovery, this joint objective requires the transmit constellation to preserve both computation-friendly superposition and sufficient geometric separability among different active-device patterns. To address this challenge, we propose a structured constellation design framework based on numerical gradient lines (NGLs), where constellation points are constrained to lie on a family of parallel lines indexed by numerical labels. We formulate a minimum-distance-based optimization problem over task-relevant superposition patterns under common NGL constraints and device power constraints. By exploiting the NGL structure, we derive a lower-dimensional scalar reformulation and further develop a low-complexity separable design. This structure orthogonally separates aggregation and active-set information, enabling closed-form finite-sum characterization of the aggregation mean-squared error (MSE) and exact active-set accuracy. The remaining scalar design parameter is adapted to the operating SNR to balance aggregation accuracy and active-set reliability. We further establish an information-theoretic necessary condition that explicitly relates exact active-set reliability, end-to-end aggregation distortion, and the information carried by the noisy finite-alphabet superposition. Numerical results validate the analytical characterization and quantify the aggregation–identification tradeoff under different system sizes and operating conditions.

Förderung

  • Swedish Foundation for Strategic Research (SSF) through the Software Artificial Intelligence for Communications (SAICOM) project
  • Knut and Alice Wallenberg Foundation (KAW) through the Turning the Air into an AI Computer (TAIRCOMP) project
  • Digital Futures, and in part by Vinnova (Sweden’s Innovation Agency) through the SYNCRO6G project

Zitieren

APA

Li, Z., & Fischione, C. (2026). Constellation Design for Active-Set-Aware Digital Over-the-Air Computation. IEEE Transactions on Communications, 74, 15703-15718. https://doi.org/10.1109/tcomm.2026.3745138

Vancouver

Li Z, Fischione C. Constellation Design for Active-Set-Aware Digital Over-the-Air Computation. IEEE Transactions on Communications. 2026;74:15703-15718. doi:10.1109/tcomm.2026.3745138

BibTeX
@article{Li2026,
  author = {Li, Zeyang and Fischione, Carlo},
  title = {Constellation Design for Active-Set-Aware Digital Over-the-Air Computation},
  journal = {IEEE Transactions on Communications},
  year = {2026},
  volume = {74},
  pages = {15703--15718},
  doi = {10.1109/tcomm.2026.3745138}
}
Kostenlos starten →