EUVIMEDEUVIMED entdecken
Crossref

SSA-Net: Spatio–Spectral Attention Network-Based Wideband Spectrum Sensing Using Sub-Nyquist Sparse Arrays

Dawei Gao, He Yan, Qinghua Guo, Bo Zhao, Yaxing Yue, Liyang Chen, Guisheng Liao

IEEE Transactions on Aerospace and Electronic Systems · 2026 · 62 · 19375-19389

DOI: 10.1109/taes.2026.3742706 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Aerospace and Electronic Systems, 2026. Verfasst von Dawei Gao und 6 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben. Gefördert von: National Natural Science Foundation of China, State Key Laboratory of Opto-Electronic Information Acquisition and Protection Technology, Anhui University, Fundamental Research Funds for the Central Universities.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This article addresses wideband spectrum sensing for cognitive receivers employing joint spatial and temporal sub-Nyquist sampling, where sparse antenna arrays and multicoset sampling introduce coupled spatio–spectral ambiguities. The ambiguities make it difficult for conventional reconstruction- or threshold-based methods to distinguish true frequency signatures from aliased components, especially under low signal-to-noise ratio (SNR) and high spectrum occupancy. In this work, we propose a spatio–spectral attention-based neural network (SSA-Net), a reconstruction-light sensing framework that directly infers spectrum occupancy from compressed measurements. SSA-Net organizes multiantenna, multicoset data into a tensor representation and employs a dual-branch spatio–spectral attention mechanism. The spectral branch models alias-dependent long-range correlations across folded subbands, while the spatial branch exploits coherent directional signatures to suppress noise and decorrelate ambiguous sources. Their fusion enables disentanglement of aliased frequency components from spatial responses, providing robust occupancy inference without relying on prior knowledge or sparse reconstruction. Simulation results show that the SSA-Net maintains reliable detection under low SNR, dense spectrum occupancy, and limited sampling channels, consistently outperforming state-of-the-art deep learning and classical methods.

Förderung

  • National Natural Science Foundation of China
  • State Key Laboratory of Opto-Electronic Information Acquisition and Protection Technology, Anhui University
  • Fundamental Research Funds for the Central Universities
  • Innovation Fund of Xidian University

Zitieren

APA

Gao, D., Yan, H., Guo, Q., Zhao, B., Yue, Y., Chen, L., & Liao, G. (2026). SSA-Net: Spatio–Spectral Attention Network-Based Wideband Spectrum Sensing Using Sub-Nyquist Sparse Arrays. IEEE Transactions on Aerospace and Electronic Systems, 62, 19375-19389. https://doi.org/10.1109/taes.2026.3742706

Vancouver

Gao D, Yan H, Guo Q, Zhao B, Yue Y, Chen L, et al. SSA-Net: Spatio–Spectral Attention Network-Based Wideband Spectrum Sensing Using Sub-Nyquist Sparse Arrays. IEEE Transactions on Aerospace and Electronic Systems. 2026;62:19375-19389. doi:10.1109/taes.2026.3742706

BibTeX
@article{Gao2026,
  author = {Gao, Dawei and Yan, He and Guo, Qinghua and Zhao, Bo and Yue, Yaxing and Chen, Liyang and Liao, Guisheng},
  title = {SSA-Net: Spatio–Spectral Attention Network-Based Wideband Spectrum Sensing Using Sub-Nyquist Sparse Arrays},
  journal = {IEEE Transactions on Aerospace and Electronic Systems},
  year = {2026},
  volume = {62},
  pages = {19375--19389},
  doi = {10.1109/taes.2026.3742706}
}
Kostenlos starten →