UDC 528.4
CSCSTI 36.23
The article presents the experimental testing results of the short-term ionospheric parameters forecasting technology based on machine learning. The basis for ionospheric parameters forecasting was data from global ionospheric maps and local ionospheric model. The model training and the formation of a data bank for training issues are considered in the paper. The authors also provide information about the software which was used for monitoring and short-term forecasting of ionospheric parameters (by using machine learning). The accuracy of the results obtained during the experimental testing of the ionospheric parameters forecasting technology is assessed. The technology considered in the article allows to be done for the selected local area: 1) creating a local model of short-term forecasting of ionospheric parameters based on machine learning; 2) modeling ionospheric parameters in a local area based on observations of the global navigation satellite system at IGS stations; 3) assessing the accuracy of short-term forecasting of the ionosphere using the input data of local modeling and the global ionospheric model. The results of the experiments showed that the shortterm ionospheric parameter forecast model trained on global ionospheric maps is capable of functioning on the basis of local ionospheric modeling data. Authors noted in the article that the standard deviation of the forecast increases by 0.468 TECU. This provides twice better accuracy as when using physical ionospheric models such as IRI-2016.
machine learning, GNSS Receivers, GNSS, monitoring of ionospheric parameters, ionospheric prediction, local ionospheric model
1. Rozhdestvenskiy D.B., Rozhdestvenskaya V.I., Smirnov V.M. i dr. Prognozirovanie parametrov ionosfery po dannym avigacionnyh sputnikovyh sistem // Naukoemkie tehnologii. 2018. T. 19. № 9. S. 21–28. DOIhttps://doi.org/10.18127/j19998465-201809-04.
2. Krasheninnikov I.V., Egorov I.B., Pavlova N.M. Effektivnost' prognozirovaniya prohozhdeniya radiovoln v ionosfere na osnove ionosfernoy modeli IRI-2001 // Geomagnetizm i aeronomiya. 2008. T. 48. № 4. S. 526–533.
3. Blaunstein N.S., Ben-Shimol Y. Prediction of Operational Parameters of Radio Signals Passing a Land-Satellite Link through StormTime Ionosphere // Information and Control Systems. 2018. Vol. 1. No. 92. P. 85–95. DOIhttps://doi.org/10.15217/issnl684-8853.2018.1.85.
4. Vasenina A.A. Vliyanie urovnya solnechnoy aktivnosti i geomagnitnyh vozmuscheniy na tochnost' prognozirovaniya kriticheskoy chastoty ionosfery // Tehnika radiosvyazi. 2014. T. 3. № 23. S. 3–10.
5. Rozhdestvenskiy D.B., Telegin V.A., Rozhdestvenskaya V.A. Vydelenie i prognozirovanie vysokochastotnyh sostavlyayuschih variaciy kriticheskoy chastoty dlya sredneshirotnoy ionosfery metodami spektral'nogo analiza // Physics of Auroral Phenomena. 2018. T. 41. № 1. S. 139–142. DOIhttps://doi.org/10.25702/KSC.2588-0039.2018.41.139-142.
6. Nava B., Coisson P., Radicella S.M. A new version of the NeQuick ionosphere electron density model // Journal of Atmospheric and Solar-Terrestrial Physics. 2008. Vol. 70. No. 15. P. 1856–1862. DOIhttps://doi.org/10.1016/j.jastp.2008.01.015.
7. Breiman L. Random Forests // Machine Learning. 2001. No. 45. P. 5–32.
8. Zamogil'nyy D. Prognozirovanie polnogo elektronnogo soderzhaniya ionosfery na osnove algoritmov mashinnogo obucheniya // Naukoemkie tehnologii v kosmicheskih issledovaniyah Zemli. 2022. T. 14. № 4. S. 39–46. DOIhttps://doi.org/10.36724/2409-5419-2022-14-4-39-46.
9. Appalonov A.M., Maslennikova Yu.S. Neyrosetevoe prognozirovanie dinamiki ekvatorial'noy anomalii po dannym polnogo elektronnogo soderzhaniya ionosfery // Tehnika radiosvyazi. 2021. T. 3. № 50. S. 29–42. DOIhttps://doi.org/10.33286/2075-8693-2021-50-29-42.
10. Serebrennikova S.A. Ocenka granic lokalizacii vozmuscheniy vysokoshirotnoy ionosfery po dannym GPS/GLONASS // Physics of Auroral Phenomena. 2021. T. 44. № 1. C. 150–153. DOIhttps://doi.org/10.51981/2588-0039.2021.44.035.
11. Zhukov A.V., Yasyukevich Yu.V., Serebennikova S.A. Mashinnoe obuchenie v zadache ocenki granicy avroral'nogo ovala po kartam variaciy polnogo elektronnogo soderzhaniya // Rasprostranenie radiovoln: trudy XXVI Vserossiyskoy otkrytoy nauchnoy konferencii: v 2 t. (Kazan', 1–6 iyulya 2019 g.). Kazan': KFU, 2019. T. 1. S. 397–400.
12. Zhukov A.V., Sidorov D.N., Myl'nikova A.A. i dr. Poisk klyuchevyh upravlyayuschih parametrov dlya operativnogo prognoza polnogo elektronnogo soderzhaniya ionosfery // Sovremennye problemy distancionnogo zondirovaniya Zemli iz kosmosa. 2018. T. 15. № 3. S. 263–272. DOIhttps://doi.org/10.21046/2070-7401-2018-15-3-263-272.
13. Salimov B.G., Berngardt O.I., Hmel'nov A.E. Primenenie svertochnyh neyronnyh setey dlya prognozirovaniya kriticheskoy chastoty foF2 // Solnechno-zemnaya fizika. 2023. T. 9. № 1. S. 60–72. DOIhttps://doi.org/10.12737/szf-91202307.
14. Reda I., Andreas A. Solar position algorithm for solar radiation applications // Solar Energy. 2004. Vol. 76. No. 5. P. 577–589.
15. Antonovich K.M. Ispol'zovanie sputnikovyh radionavigacionnyh sistem v geodezii: monografiya: v 2 t. M.: Kartgeocentr, 2005. T. 1. 334 s.
