SIGNAL DEMODULATION USING CLASSICAL MACHINE LEARNING ALGORITHMS FOR THE WATTERSON MODEL
Abstract
This paper investigates the application of classical machine learning algorithms for signal demodulation in a high-frequency communication channel described by the Watterson model. The relevance of this task stems from the need to improve noise immunity during broadband data transmission under conditions of multipath propagation and Doppler distortions. The classifiers used include Random Forest, Decision Tree, and k-nearest neighbors (KNN). The models are trained on synthesized data generated in the MATLAB Simulink environment, with varying path delays (0–4 ms) and Doppler shifts (0–10 Hz). The input features comprise the coordinates of signal constellations as well as the history of received symbols, since intersymbol interference significantly affects the position of the current point. The experiments revealed a critical feature: demodulation algorithms exhibit high sensitivity to variations in Doppler shift. A deviation of just 0.1 Hz from the training set parameters alters the structure of the signal constellation, increasing the bit error rate (BER) to 0.5. When the Doppler parameters align with the training grid, the Random Forest algorithm demonstrates the best performance, achieving a BER <0.01 on the full dataset. To address this sensitivity, a local training method with a reduced Doppler shift sampling step (down to 0.1 Hz) is proposed, enabling a BER ≤0.01 for Random Forest. Due to the exponential growth of the training dataset size (up to 1000 GB), data reduction techniques were developed: excluding time parameters and sampling data based on possible combinations of preceding bits. This reduced the data volume by a factor of 31.5 while maintaining a BER <10% for Random Forest with a training step of 0.2 Hz. The study concludes that the Random Forest classifier is suitable for demodulation in HF channels and highlights the necessity of adapting the training step to the required Doppler shift accuracy.
##article.references##
1. Watterson C.C., Juroshek J.R., Bensema W.D. Experimental confirmation of an HF channel model, IEEE(R) Trans. commun. Technol., 1970, No. Dec. 6, Vol. COM-18, pp. 792-803.
2. O’shea T., Hoydis J. An introduction to deep learning for the physical layer, IEEE Transactions on Cognitive Communications and Networking, 2017, Vol. 3, No. 4, pp. 563-575.
3. Song M., Song X., Qin K. Signal Detection and Demodulation Algorithm Based on Deep Learning in Communication Network, 2024 3rd International Conference on Artificial Intelligence and Autonomous Robot Systems (AIARS), IEEE, 2024, pp. 461-466.
4. Fang L., Wu L. Deep learning detection method for signal demodulation in short range multipath channe, 2017 IEEE 2nd International Conference on Opto-Electronic Information Processing (ICOIP). IEEE, 2017, pp. 16-20.
5. Erokhin S.D., Chadov T.A., Zhuravlev A.P. Osobennosti formirovaniya obuchayushchey i testovoy vyborok dlya algoritmov mashinnogo obucheniya, ispol'zuemykh pri optimizatsii raboty priemnykh ustroystv [Features of the formation of training and test samples for machine learning algorithms used in optimizing the operation of receiving devices], DSPA: Voprosy primeneniya tsifrovoy obrabotki signalov [DSPA: Application Issues of Digital Signal Processing], 2020, Vol. 10, No. 1, pp. 9-14.
6. Malygin I.V. i dr. Primenenie metodov mashinnogo obucheniya dlya klassifikatsii radiosignalov [Appli-cation of machine learning methods for classification of radio signals], Tr. MAI [Proceedings of MAI], 2017, No. 96, pp. 15.
7. Chirkov O.N., Pirogov A.A. Primenenie algoritmov mashinnogo obucheniya v zadache otsenki be-sprovodnogo kanala svyazi s OFDM [Application of machine learning algorithms in the problem of as-sessing a wireless communication channel with OFDM], Vestnik Voronezhskogo gosudarstvennogo tekhnicheskogo universiteta [Bulletin of the Voronezh State Technical University], 2023, Vol. 19,
No. 6, pp. 164-169.
8. Kolesnikov A.A. i dr. Ispol'zovanie tekhnologiy mashinnogo obucheniya pri reshenii geoinfor-matsionnykh zadach [Using machine learning technologies in solving geoinformation problems], Inter-Karto. InterGis [InterCarto. InterGis], 2018, Vol. 24, No. 2, pp. 371-384.
9. Sinyutin S.A., Sinyutin E.S., Yartsev A.V. Issledovanie ustoychivosti modeli mashinnogo obucheniya dlya obrabotki signala pribora orientatsii po Zemle [Study of the stability of a machine learning model for processing the signal of an Earth orientation device], Matematicheskoe modelirovanie [Mathematical Modeling], 2023, Vol. 35, No. 6, pp. 3-13.
10. Bulat I.V. i dr. Sovremennoe sostoyanie metodov mashinnogo obucheniya v prilozheniyakh obrabotki odnomernykh signalov [Current state of machine learning methods in one-dimensional signal processing applications], Elektronika, fotonika i kiberfizicheskie sistemy [Electronics, Photonics and Cyber-Physical Systems], 2022, Vol. 2, No. 3, pp. 86-95.
11. Ivanov D.V. i dr. Novye vozmozhnosti sistem shirokopolosnoy kognitivnoy svyazi, rabotayushchikh v ionosfernykh KV-radiokanalakh s vnutrimodovoy dispersiey [New capabilities of broadband cognitive communication systems operating in ionospheric HF radio channels with intramode dispersion], Radio-tekhnika [Radio Engineering], 2022, Vol. 86, No. 11, pp. 162-177.
12. Konkin N.A. Metodika i algoritm opredeleniya periodov operativnogo prognozirovaniya dinamiki maksimal'no primenimykh chastot KV-svyazi na osnove algoritma mashinnogo obucheniya XGBoost [Methodology and algorithm for determining the periods of operational forecasting of the dynamics of the maximum applicable frequencies of HF communication based on the XGBoost machine learning al-gorithm], Vestnik Povolzhskogo gosudarstvennogo tekhnologicheskogo universiteta. Seriya: radio-tekhnicheskie i infokommunikatsionnye sistemy [Bulletin of the Volga Region State Technological Uni-versity. Series: Radio Engineering and Infocommunication Systems], 2022, No. 3, pp. 6-16.
13. Ivanov D.V. i dr. Obespechenie predel'noy shirokopolosnosti sistem kognitivnoy nazemnoy i trasion-osfernoy radiosvyazi, rabotayushchikh v usloviyakh dispersionnykh iskazheniy [Ensuring the maximum bandwidth of cognitive terrestrial and transionospheric radio communication systems operating under conditions of dispersion distortions], Rasprostranenie radiovoln [Radio Wave Propagation], 2023, pp. 5-13.
14. Kazantseva A.A. Primenenie metodov glubokogo mashinnogo obucheniya dlya prognozirovaniya dostupnosti radiokanalov KV-svyazi [Application of deep machine learning methods for predicting the availability of HF radio channels], Sb. dokladov I Vserossiyskoy molodezhnoy nauchnoy shkoly-konferentsii [Collection of reports of the 1st All-Russian Youth Scientific School-Conference], 2024, pp. 107-110.
15. Kirillov A.S. Algoritmy demodulyatsii dlya kanala svyazi Vattersona s izvestnymi parametrami [Demod-ulation algorithms for the Watterson communication channel with known parameters], Radiotekhnika [Radio Engineering], 2025, Vol. 89, No. 4. pp. 43-50.
16. Kirillov A.S. Machine Learning Algorithms for Signal Demodulation and Determination of HF Commu-nication Channel Parameters, IEEE Radiation and Scattering of Electromagnetic Waves (RSEMW), 2023, pp. 268-271.
17. Kirillov A.S. Opredelenie parametrov modeli Vattersona s pomoshch'yu adaptivnykh fil'trov [Determin-ing the parameters of the Watterson model using adaptive filters], Radiotekhnika [Radio Engineering], 2024, No. 10, pp. 5-13.
18. Kirillov A.S. Algoritmy mashinnogo obucheniya priem signala tsifrovykh radioliniy [Machine learning algorithms for receiving digital radio lines], Radiolokatsiya i radiosvyaz' [Radar and Radio Communica-tions], 2022, No. 15, pp. 265.
19. Recommendation ITU-R F.1487, Testing of HF modems with bandwidths of up to about 12 kHz using ionospheric channel simulators, 2000.
20. Recommendation ITU-R F.520, Use of high frequency ionospheric channel simulators, 1992.
