EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Deep Learning-Based Automatic Modulation Classification for OFDM Signals: From Synthetic Training to OTA Evaluation

Raluca Nelega iD, Mate-Marton Mezei iD, Zsolt Alfred Polgar iD, Gergo Kovacs, Emanuel Puschita iD

Sensors · 08.05.2026 · 26 (10) · 2945

DOI: 10.3390/s26102945 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 08.05.2026. Verfasst von Raluca Nelega und 4 weiteren Autor:innen. Beteiligte Einrichtung: National Institute for Research and Development of Isotopic and Molecular Technologies, 67-103 Donat Street, 400293 Cluj-Napoca, Romania. Die Arbeit nennt 27 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Research, Innovation and Digitization, European Regional Development Fund.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

To address the growing congestion of the radio frequency (RF) spectrum, Cognitive Radio (CR) systems employ Automatic Modulation Classification (AMC) to dynamically optimize spectrum utilization without introducing protocol overhead. In modern Orthogonal Frequency Division Multiplexing (OFDM) standards, effective AMC requires advanced signal-processing techniques capable of accurately identifying modulation schemes under dynamic channel conditions. Therefore, maintaining robust performance under realistic environments remains a fundamental challenge. This paper evaluates how dataset scale, synthetic impairments, and hardware-induced signal impairments affect the cross-domain generalization of a Convolutional Neural Network (CNN) architecture for OFDM Automatic Modulation Classification (AMC), using 2D amplitude-phase histograms for signal representation. To assess these effects, the CNN is trained on five distinct datasets, encompassing both synthetically generated signals with varying scales and synchronization impairments, as well as a conducted hardware dataset. The cross-domain generalization of the trained models is assessed by evaluating them on a completely unseen indoor Over-The-Air (OTA) dataset collected across 13 distinct positions. Statistical analysis demonstrates that the large-scale synchronization-impaired synthetic dataset achieves the best generalization performance, reaching a mean indoor OTA accuracy of 93.36% and outperforming the limited-size conducted hardware dataset. Overall, this study demonstrates the critical role of data-generation strategies and establishes a robust baseline for achieving reliable cross-domain generalization of CNN-based AMC.

Schlagwörter

Stichwörter der Autor:innen

CNNautomatic modulation classificationcross-domain generalizationdeep learningsignal processing

Autor:innen und Institutionen

  • National Institute for Research and Development of Isotopic and Molecular Technologies, 67-103 Donat Street, 400293 Cluj-Napoca, Romania.
  • Communications Department, Technical University of Cluj-Napoca, 26-28 George Baritiu Street, 400027 Cluj-Napoca, Romania.

ORCID-Profile: Raluca Nelega, Mate-Marton Mezei, Zsolt Alfred Polgar, Emanuel Puschita

Förderung

  • Ministry of Research, Innovation and Digitization
  • European Regional Development Fund

Zitieren

APA

Nelega, R., Mezei, M. M., Polgar, Z. A., Kovacs, G., & Puschita, E. (2026). Deep Learning-Based Automatic Modulation Classification for OFDM Signals: From Synthetic Training to OTA Evaluation. Sensors, 26(10), 2945. https://doi.org/10.3390/s26102945

Vancouver

Nelega R, Mezei MM, Polgar ZA, Kovacs G, Puschita E. Deep Learning-Based Automatic Modulation Classification for OFDM Signals: From Synthetic Training to OTA Evaluation. Sensors. 2026;26(10):2945. doi:10.3390/s26102945

BibTeX
@article{Nelega2026,
  author = {Nelega, Raluca and Mezei, Mate-Marton and Polgar, Zsolt Alfred and Kovacs, Gergo and Puschita, Emanuel},
  title = {Deep Learning-Based Automatic Modulation Classification for OFDM Signals: From Synthetic Training to OTA Evaluation},
  journal = {Sensors},
  year = {2026},
  volume = {26},
  number = {10},
  pages = {2945},
  doi = {10.3390/s26102945}
}
Kostenlos starten →