EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Reconstruction of SINR Maps From Sparse Measurements Using Group Equivariant Non-Expansive Operators

Lorenzo Mario Amorosa, Francesco Conti, Nicola Quercioli, Flavio Zabini, Tayebeh Lotfi Mahyari, Yiqun Ge, Patrizio Frosini

IEEE Transactions on Machine Learning in Communications and Networking · 2026 · 4 · 1353-1370

DOI: 10.1109/tmlcn.2026.3716686 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Machine Learning in Communications and Networking, 2026. Verfasst von Lorenzo Mario Amorosa und 6 weiteren Autor:innen. Die Arbeit nennt 42 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Framework of the National Wireless Communication Laboratory (CNIT) National Laboratory WiLab, WiLab-Huawei Joint Innovation Center, INdAM-GNSAGA, in part by the COST Action CaLISTA.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

As sixth generation (6G) wireless networks evolve, accurate signal-to-interference-noise ratio (SINR) maps are becoming increasingly critical for effective resource management and optimization. However, acquiring such maps at high resolution is often cost-prohibitive, creating a severe data scarcity challenge. This necessitates machine learning (ML) approaches capable of robustly reconstructing the full map from extremely sparse measurements. To address this, we introduce a novel reconstruction framework based on Group Equivariant Non-Expansive Operators (GENEOs). Unlike data-hungry ML models, GENEOs are low-complexity operators that embed domain-specific geometric priors, such as translation and rotational equivariance, directly into their structure. This provides a strong inductive bias, enabling effective reconstruction from very few samples. Our key insight is that for network management, preserving the topological structure of the SINR map, such as the geometry of coverage holes and interference patterns, is often more critical than minimizing pixel-wise error. We validate our approach on realistic ray-tracing-based urban scenarios, evaluating performance with both traditional statistical metrics (mean squared error (MSE)) and, crucially, a topological metric (1-Wasserstein distance). Results show that our method consistently achieves superior statistical and topological accuracy across diverse urban scenarios. Compared to the best-performing baselines, GENEO reduces MSE up to 45% and decreases the 1-Wasserstein distance up to 54%. Crucially, these substantial performance gains are maintained even under the most extreme tested conditions, such as a 1% sampling rate with a 30% measurement error, and when measurements are highly spatially biased. This demonstrates the practical advantage of GENEOs for creating structurally accurate SINR maps that are more reliable for downstream network optimization tasks.

Förderung

  • Framework of the National Wireless Communication Laboratory (CNIT) National Laboratory WiLab
  • WiLab-Huawei Joint Innovation Center
  • INdAM-GNSAGA, in part by the COST Action CaLISTA
  • HORIZON Research and Innovation Action PANDORA

Zitieren

APA

Amorosa, L. M., Conti, F., Quercioli, N., Zabini, F., Mahyari, T. L., Ge, Y., & Frosini, P. (2026). Reconstruction of SINR Maps From Sparse Measurements Using Group Equivariant Non-Expansive Operators. IEEE Transactions on Machine Learning in Communications and Networking, 4, 1353-1370. https://doi.org/10.1109/tmlcn.2026.3716686

Vancouver

Amorosa LM, Conti F, Quercioli N, Zabini F, Mahyari TL, Ge Y, et al. Reconstruction of SINR Maps From Sparse Measurements Using Group Equivariant Non-Expansive Operators. IEEE Transactions on Machine Learning in Communications and Networking. 2026;4:1353-1370. doi:10.1109/tmlcn.2026.3716686

BibTeX
@article{Amorosa2026,
  author = {Amorosa, Lorenzo Mario and Conti, Francesco and Quercioli, Nicola and Zabini, Flavio and Mahyari, Tayebeh Lotfi and Ge, Yiqun and Frosini, Patrizio},
  title = {Reconstruction of SINR Maps From Sparse Measurements Using Group Equivariant Non-Expansive Operators},
  journal = {IEEE Transactions on Machine Learning in Communications and Networking},
  year = {2026},
  volume = {4},
  pages = {1353--1370},
  doi = {10.1109/tmlcn.2026.3716686}
}
Kostenlos starten →