A.A. Petrov1, O.V. Druzhinina2, O.N. Masina3

1, 3 Bunin Yelets State University (Yelets, Russia)

2 FRС «Computer Science and Control» of RAS (Moscow, Russia)

The study of the propagation of electromagnetic waves is one of the most relevant scientific directions. The range of important problems includes modeling the mechanism of propagation of Wi-Fi radio waves in inhomogeneous media. The development of an approach to solving this problem based on the construction and implementation of an artificial neural network is promising and allows to take into account the data of field experiments in the presence of uncertainties. The objectives of the article include the justification of the choice of neural network architecture and the development of algorithms for data preparation and machine learning for modeling the propagation of electromagnetic waves in inhomogeneous media. The simulation of the propagation process of electromagnetic waves is carried out. As an example, the process of propagation of Wi-Fi radio waves is considered. Algorithms of data preparation for neural network modeling are implemented. A machine learning algorithm with reinforcement has been developed. Specialized software is created, a series of computational experiments has been performed and their interpretation is given. The proposed models and the obtained results can be used in the tasks of intelligent forecasting and monitoring of the propagation of electromagnetic waves when improving digital means of communication.

Petrov A.A., Druzhinina O.V., Masina O.N. Approach to neural network modeling of the propagation of electromagnetic waves. Electromagnetic waves and electronic systems. 2022. V. 27. № 6. P. 53−58. DOI: https://doi.org/10.18127/j15604128-202206-07 (in Russian)

1. Utts V.L. Issledovanie poter pri rasprostranenii radiosignala sotovoi svyazi na osnove statisticheskikh modelei. Vestnik Baltiiskogo Federalnogo universiteta im. I. Kanta. 2011. № 5. S. 44−49. (in Russian)
2. Milyutin E.R., Vasilenko G.O., Sivers M.A. i dr. Metody rascheta polya v sistemakh svyazi detsimetrovogo diapazona. SPb.: Triada. 2003. (in Russian)
3. Hata M. Empirical formula for propagation loss in land mobile radio services. IEEE Transactions on Vehicular Technology.1980. V. VT-29. № 3. P. 317−325.
4. Grishko A.K., Goryachev N.V., Trusov V.A. i dr. Analiz matematicheskikh modelei dlya rascheta parametrov radiotrass. Trudy Mezhdunar. simpoziuma «Nadezhnost i kachestvo». 2018. T. 2. S. 361−363. (in Russian)
5. Panchenko V.E., Erokhin G.A., Gainutdinov T.A. i dr. Sochetanie statisticheskikh i deterministskikh metodov rascheta radiopolya v gorodskikh usloviyakh. Elektrosvyaz. 1998. № 4. S. 31−33. (in Russian)
6. Molchanov S.V., Zakharov A.I. Primenenie iskusstvennykh neironnykh setei s tselyu prognozirovaniya kharakteristik rasprostraneniya radiosignala. Vestnik Baltiiskogo Federalnogo universiteta im. I. Kanta. 2014. № 4. S. 100−105.
7. Wilson H. Artificial intelligence. Grey House Publishing. 2018. (in Russian)
8. Nikolenko S.I., Kadurin A.A., Arkhangelskaya E.V. Glubokoe obuchenie. SPb.: Piter. 2018. (in Russian)
9. Petrov A.A. Struktura programmnogo kompleksa dlya modelirovaniya tekhnicheskikh sistem v usloviyakh pereklyucheniya rezhimov raboty. Elektromagnitnye volny i elektronnye sistemy. 2018. T. 23. № 4. S. 61−64. (in Russian)
10. Petrov A.A., Druzhinina O.V., Masina O.N. Razvitie algoritmicheskogo obespecheniya dlya modelirovaniya nelineinykh upravlyaemykh sistem s pereklyucheniyami. Nelineinyi mir. 2022. T. 20. № 1. S. 5−13. (in Russian)
11. Chernomordov S.V., Druzhinina O.V., Masina O.N., Petrov A.A. Primenenie metodov mashinnogo obucheniya v zadachakh neirosetevogo modelirovaniya upravlyaemykh tekhnicheskikh sistem. Neirokompyutery: razrabotka, primenenie. 2022. T. 24. № 1. S. 25−35. (in Russian)
12. Roth Scott D. Ray Casting for Modeling Solids. Computer Graphics and Image Processing. 1982. V. 18. P. 109−144.