NEURAL NETWORK METHOD FOR INVESTIGATION SPECTRAL CHARACTERISTICS
DOI:
https://doi.org/10.34169/2414-0651.2025.2(46).35-44Keywords:
power, approach, validation, testing, model, system, feature, classification, radio signals, amplitude, phaseAbstract
This article considers the problem of radio signal classification based on spectral features formed from complex low-frequency signal samples (in-phase and quadrature components). The main goal of the research is to build a single machine learning model capable of effectively identifying the type of signal by its spectral characteristics. The signal is represented using the power spectral density (PSD), calculated by the Welch method, as well as additional statistical and frequency-energy features that reflect the amplitude-phase structure of the signal. The model structure is proposed and the processes of its training, validation and testing are implemented. An analysis of the influence of spectral decomposition parameters on classification quality is conducted. The experimental results demonstrate that the combined use of spectral and statistical features allows achieving high accuracy in recognizing various types of radio signals. The proposed approach can be applied in practical systems for automatic radio frequency spectrum analysis and signal detection in complex electromagnetic environments.
Downloads
References
Kumar, S., Ahmad, I., Höyhtyä, M., Khan, S. & Gurtov, A. Deep Learning Frameworks for Cognitive Radio Networks: Review and Open Research Challenges. Available at: https://arxiv.org/pdf/2410.23949.
Kim, S., Kim, J., Doan, V. & Kim, D. Lightweight Deep Learning Model for Automatic Modulation Classification in Cognitive Radio Networks. Available at: https://www.researchgate.net/publication/346433098_Lightweight _Deep_Learning_Model_for_Automatic_Modulation_Classification_in_Cognitive_Radio_Networks.
Fekry, O., Abdalla, M. & Elsayed, A. Deep Learning-Based Automatic Modulation Classification Using Robust CNN Architecture for Cognitive Radio Networks. Available at: https://www.researchgate.net/publication/ 346433098_Lightweight_Deep_Learning_Model_for_Automatic_Modulation_Classification_in_Cognitive_Radio_Networks.
Slyusar, V.I. & Masesov, N.A. (2008). The limits of correcting quadrature imbalance using additional strobing of ADC samples. In: Proc. of the 2nd Intern. Scientific and Technical Conf. «Problems of Telecommunications». K.: Inst. of Telecommunication Systems of NTUU «KPI». Pp. 198—199.
Slyusar, V.I., Serdyuk, P.E. & Zhyvylo, E.A. (2011). The influence of digital I/Q demodulation on OFDM signals. In: Proc. of the Intern. Scientific and Technical Conf. «Information Systems and Technologies (IST-2011)». Nizhny Novgorod State Technical Univ. April 22. P. 45.
Slyusar, V.I. (2021). Neural network models based on tensor-matrix theory. In: «Problems of the Development of Advanced Micro- and Nanoelectronic Systems» (MES-2021). No. 2. Pp. 23—28. https://doi.org/10.31114/2078-7707-2021-2-23-28. DOI: https://doi.org/10.31114/2078-7707-2021-2-23-28
Slyusar, V.I. (2021). Multimodal Quasifractal Neural Networks. In: Proc. of the XX Intern. Scientific Conf. «Neural Network Technologies and Their Applications (NMTiZ-2021)». Kramatorsk: Donbas State Engineering Acad. December 8–9. Pp. 134—137.
Slyusar, V.I. & Bihun, N.S. A neural network for protecting UAV communication channels. In: Abstracts of the Intern. Scientific and Technical Conf. «Prospects for the Development of Armament and Military Equipment of the Land Forces». Lviv. May 17–18, 2023.
PySDR: A Guide to SDR and DSP using Python. 2022. Available at: https://pysdr.org.
SciPy Documentation – Signal Processing. Available at: https://docs.scipy.org/doc/scipy/reference/signal.html.
IQ Data Explained. PE0SAT Satellite Ground Station: [website]. Available at: https://www.pe0sat.vgnet.nl/sdr/iq-data-explained/.
IQ Signal Master Vector Signal Analysis Software. Anritsu: [website]. Available at: https://www.anritsu.com/en-us/test-measurement/products/mx280005a.
Lecture 9: Analog and Digital I/Q Modulation. Massachusetts Institute of Technology (MIT): [website]. Available at: https://web.mit.edu/6.02/www/f2006/handouts/Lec9.pdf.
IQ Signal Master™ MX280005A Vector Signal Analysis Software. Anritsu: Available at: https://dl.cdn-anritsu.com/en-us/test-measurement/files/Brochures-Datasheets-Catalogs/Brochure/11410-02844F.pdf.
West, N. & O'Shea, T. End-to-End Learning From Spectrum Data: A Deep Learning Approach for Wireless Signal Identification in Spectrum Monitoring Applications. ResearchGate: [website]. Available at: https://www.researchgate.net/publication/321761024_End-to-End_Learning_From_Spectrum_Data_A_Deep _Learning_Approach_for_Wireless_Signal_Identification_in_Spectrum_Monitoring_Applications.
O’Shea, T.J., West, N. & Clancy, T.C. (2019). Classification of Radio Signals and HF Transmission Modes with Deep Learning. Available at: https://arxiv.org/pdf/1906.04459.
Data Analysis in Python. ReadTheDocs: [website]. Available at: https://dataanalysispython.readthedocs.io/_/ downloads/en/latest/pdf/.
Lyons, R. Quadrature Signals: Complex, But Not Complicated. IEEE.li: [website]. Available at: https://www.ieee.li/pdf/essay/quadrature_signals.pdf.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2025 Вадим Слюсар,Вадим Козлов,Денис Козлов
This work is licensed under a Creative Commons Attribution 4.0 International License.
