V.A. Ivanov1, N.V. Ryabova2, N.A. Konkin3, M.I. Bastrakova4, A.A. Chernov5

1–5 Volga State University of Technology (Yoshkar-Ola, Russia)

1 IvanovVA@volgatech.net, 2 RyabovaNV@volgatech.net, 3 KonkinNA@volgatech.net,
4 BastrakovaMI@volgatech.net, 5 ChernovAA@volgatech.net

The use of machine learning methods for predicting the optimal operating frequency (ORF) of short-wave (SW) communication lines makes it possible to increase accuracy compared to other forecasting methods and thereby ensure the reliability of information transmission in conditions of interference, signal fading, and in the absence of diagnostic equipment. Methods have been developed for assessing the availability of the HF channel at the optimal operating frequency, assigned according to the forecast and determined by a sensor with verification on a mid-latitude experimental radio link with a length of 2.6 thousand km. To predict the MUF, the ensemble machine learning model Comb_Voting_Huber_p–10_mae was used, which showed the best results according to the criterion of minimum deviation from experimental data (3–5%) on average for the entire study period. Based on the predicted and experimentally obtained ORF values, the availability of such radio channels was determined and the accuracy of the results obtained in the conditions of their seasonal and daily variations was assessed. It was found that when using machine learning, the average annual forecast availability is 96.756%, and the experimental one is 99.8%. Experimental studies revealed the limits of variability in channel availability values, which turned out to be maximum in winter and minimum in summer. Periods of increased variability correlated with variations in the level of geomagnetic activity and ionospheric disturbances. It has been established that even in conditions of a disturbed ionosphere, the predictive ensemble model of machine learning shows stability and suitability for practical use in adaptive HF radio communication systems.

Ivanov V.A., Ryabova N.V., Konkin N.A., Bastrakova M.I., Chernov A.A. Analysis of the availability and accuracy of its assessment for the HF radio channel on the predicted ORF determined by the machine learning method. Electromagnetic waves and electronic systems. 2026. V. 31. № 2. P. 78−87. DOI: https://doi.org/10.18127/j15604128-202602-09 (in Russian)

1. Ivanov D.V., Ivanov V.A., Ryabova N.V., Elsukov A.A., Konkin N.A. Active sensor with remote control for diagnostics of broadband ionospheric radio channels with OFDM-BPSK-signals. Radiotekhnika. 2022. V. 86. № 12. P. 90−104. DOI doi.org/10.18127/j00338486-202212-08. (in Russian)
2. Ivanov V.A., Ivanov D.V., Ryabova N.V., Chernov A.A. Multidimensional high-frequency radio channel and experimental studies of its main characteristics. Electromagnetic waves and electronic systems. 2013. V. 18. № 8. P. 40–48. (in Russian)
3. Bilitza D., Pezzopane M., Truhlik V., Altadill D., Reinisch B.W., Pignalberi A. The International Reference Ionosphere Model: A Review and Description of an Ionospheric Benchmark. Reviews of Geophysics. 2022. V. 60. № 4. P. e2022RG000792. DOI 10.1029/ 2022RG000792.
4. Tsagouri I.I., Temens D.R., Belehaki A., Shim J.-S., Hoque M., Nykiel G., Borries C., Morozova A., Barata T., Miloch W.J. Ionosphere Variability II: Advances in theory and modeling. Advances in Space Research. 2023. DOI 10.1016/j.asr.2023.07.056.
5. Lv Z.Y., Zhang K.T. Multi-hop HF Radio Propagation. International Conference on Information Science, Parallel and Distributed Systems (ISPDS). Xi'an, China. 2020. P. 119–127. DOI 10.1109/ISPDS51347.2020.00032.
6. Ryabova N.V., Konkin N.A., Chernov A.A., Ivanova N.V. Comparative analysis of regression and ensemble methods of machine learning in order to predict the MHR of ionospheric radio links. Proceedings of the XXV International Scientific and Technical Conference "Telecommunications: theory and Technology" TTT-2023. Kazan: Kazan National Research Technical University named after A.N. Tupolev-KAI. 2023. P. 18–20. (in Russian)
7. Zheng A., Casari A. Feature Engineering for Machine Learning. Sebastopol, CA: O’Reilly Media. 2018. 219 p.
8. Ivanov V.A., Ivanov D.V., Mikheeva N.N., Ryabova M.I. Dispersion distortions of system characteristics of broadband ionospheric radio channels: monograph. Yoshkar-Ola. 2015. 160 p. (in Russian)
9. Ivanov D.V., Ivanov V.A., Ryabova N.V., Bel'gibaev R.R., Chernjad'ev A.V. Monitoring of Interference Spectrum and Availability of HF Radio Channels of 3…24 KHz Width. Vestnik of Volga State University of Technology. Ser.: Radio Engineering and Infocommunication Systems. 2022. № 1(53). P. 21–32. DOI 10.25686/2306-2819.2022.1.21. (in Russian)
10. ITU-R F.763-4. Radio-frequency channel noise and system performance for HF sky-wave communication systems. Geneva: ITU. 1999. 34 p.