Prediction of crosstalk in aircraft cable communication lines using artificial intelligence technologies


Аuthors

Amirkhanov A. A.*, Gaynutdinov R. R.**

Kazan National Research Technical University named after A.N. Tupolev, Kazan, Russia

*e-mail: reimeartorias@mail.ru
**e-mail: emc-kai@mail.ru

Abstract

Abstract. Currently, the study of crosstalk in cable communication lines is carried out using computer modeling or experimental studies. This paper proposes an innovative approach to predicting cross-electromagnetic interference in cable communication lines using artificial intelligence technologies. The development and training of a fully connected direct propagation neural network for predicting crosstalk between cables in the frequency domain is presented. The proposed approach leverages the capabilities of artificial neural networks to learn complex patterns in the data and make accurate predictions. The neural network architecture is designed to process large datasets quickly and efficiently, making it an ideal solution for crosstalk analysis in cable communication lines. A comprehensive comparative analysis of the learning results of an artificial neural network with different values of hyperparameters is presented. The hyperparameters of the neural network structure have been determined to ensure the lowest prediction error, resulting in a highly accurate and reliable model. The paper presents practical examples of the operation of such a neural network, demonstrating its ability to accurately predict crosstalk values for different frequency ranges and cable configurations. The proposed approach has the potential to significantly reduce the time and cost of crosstalk analysis in cable communication lines, making it a valuable tool for the telecommunications industry. When analyzing the data, it was determined that the neural network is capable of predicting crosstalk in cable communication lines with sufficiently high accuracy. Thus, the results of comparing the levels of electromagnetic interference obtained using computer modeling with the results of the ANN show an average absolute error not exceeding 6.77%. The trained neural network can be used to solve the problem of tracing cables in technical objects, including aircraft. In this case, cross-electromagnetic interference calculated using the presented neural network can be used as a tracing criterion.

Keywords:

neural network, crosstalk, forecasting, electromagnetic compatibility, computer modeling, aircraft

References

  1. Kirillov V.Yu., Marchenko M.V., Tomilin M.M. Elektromagnitnaya sovmestimost' bortovoi kabel'noi seti letatel'nykh apparatov (Electromagnetic compatibility of the on-board cable network of aircraft). Moscow: MAI Publ., 2014. 172 p.
  2. Kechiev L.N., Balyuk N.V. Zarubezhnye voennye standarty v oblasti EMS (Foreign military standards in the field of EMS). Moscow: Grifon Publ., 2014. 448 p.
  3. Gainutdinov R.R., Chermoshentsev S.F. Methodology for ensuring intra-system electromagnetic compatibility of on-board equipment of unmanned aerial vehicles. Izvestiya vysshikh uchebnykh zavedenii. Aviatsionnaya tekhnika. 2016. No. 4. P. 155-160. (In Russ.)
  4. Gainutdinov R.R., Chermoshentsev S.F. Forecasting of cross-electromagnetic interference in cable communication lines of an unmanned aerial vehicle. Vestnik Kazanskogo gosudarstvennogo tekhnicheskogo universiteta im. A.N. Tupoleva. 2014. No. 3. P. 178-182. (In Russ.)
  5. Gainutdinov R.R., Chermoshentsev S.F. Prediction of electromagnetic interference in communication lines of interblock aircraft with microsecond pulsed electromagnetic interactions. Vestnik Kazanskogo gosudarstvennogo tekhnicheskogo universiteta im. A.N. Tupoleva. 2014. No. 2. P. 203-208. (In Russ.)
  6. Zabolotskii A.M., Gazizov T.R., Kalimulin I.F. Novye resheniya dlya obespecheniya elektromagnitnoi sovmestimosti bortovoi radioelektronnoi apparatury kosmicheskogo apparata: monografiya (New solutions for ensuring electromagnetic compatibility of on-board electronic equipment of the spacecraft: monograph). Tomsk: Tomskii gosudarstvennyi universitet sistem upravleniya i radioelektroniki Publ., 2016. 288 p.
  7. Nguen V.T., Kirillov V.Yu. Designing routes of electrical harnesses of the on-board network taking into account electromagnetic compatibility. Tekhnologii elektromagnitnoi sovmestimosti. 2020. No. 2 (73). P. 29-35. (In Russ.)
  8. Gainutdinov R.R., Chermoshentsev S.F. Electromagnetic compatibility of promising aviation complexes. Tekhnologii elektromagnitnoi sovmestimosti. 2018. No. 2 (65). P. 62-78. (In Russ.)
  9. Klykov A.V., Kirillov V.Yu. Possibilities of computer modeling in solving problems of electromagnetic compatibility of on–board aircraft networks. Trudy MAI. 2012. No. 57. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=30760
  10. Kuksenko S.P., Zabolotskii A.M. Elektromagnitnaya sovmestimost': modelirovanie i obespechenie (Electromagnetic compatibility: modeling and support). Tomsk: Tomskii gosudarstvennyi universitet sistem upravleniya i radioelektroniki Publ., 2017. 96 p.
  11. Polyanskii A.V., Smykova N.N., Nikonorova L.I. Neural networks today and development prospects. Nauka i Obrazovanie. 2023. V. 6, No. 2. (In Russ.). URL: http://opusmgau.ru/index.php/see/article/view/5814/5832
  12. Mansurov A.S. Introduction of an inter–cable novodok for ANSYS MAXWELL 2D software. Aktual'nye problemy gumanitarnykh i estestvennykh nauk. 2014. No. 5-1. P. 54-59. (In Russ.)
  13. Amirkhanov A.A. Analysis of crosstalk in cable lines of a structured cable system. Vserossiiskaya molodezhnaya nauchnaya konferentsiya «Budushchee nauki: vzglyad molodykh uchenykh na innovatsionnoe razvitie obshchestva»: sbornik nauchnykh statei. Kursk: Universitetskaya kniga Publ., 2023. V. 2, P. 378-381.
  14. Voronov K.E., Grigor'ev D.P., Telegin A.M. Application of a direct propagation neural network for localization of the impact site of microparticles on the surface of a spacecraft. Trudy MAI. 2021. No. 118. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=158245. DOI: 10.34759/trd-2021-118-10
  15. Stepanenko A.S., Shchegol'kov A.S. Neural networks as a forecasting tool in civil aviation. XII Mezhdunarodnaya nauchno-prakticheskaya konferentsiya, posvyashchennaya prazdnovaniyu 100-letiya otechestvennoi grazhdanskoi aviatsii «Aktual'nye problemy i perspektivy razvitiya grazhdanskoi aviatsii»: sbornik trudov. Moscow: Moskovskii gosudarstvennyi tekhnicheskii universitet grazhdanskoi aviatsii Publ., 2023. P. 211-219.
  16. Kozadaev A.S. Principles of artificial neural network implementations. Vestnik Tambovskogo universiteta. Seriya: Estestvennye i tekhnicheskie nauki. 2007. V. 12. No. 1. P. 108-110. (In Russ.)
  17. Kozlov D.S., Tyumentsev Yu.V. Neural network methods for detecting failures of sensors and drives of an aircraft. Trudy MAI. 2012. No. 52. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=29421
  18. Starovoitov V.V., Golub Yu.I. Systematization of data in machine learning. Informatika. 2021. V. 18, No. 3. P. 83-96. (In Russ.). DOI: 10.37661/1816-0301-2021-18-3-83-96
  19. Anuarbekov A.N., Iskakov K.T., Mukhametzhanova B.O. Application of convolutional neural networks: types, varieties and approaches. III Mezhdunarodnaya nauchno-prakticheskaya konferentsiya «Razvitie sovremennoi nauki: opyt teoreticheskogo i empiricheskogo analiza»: sbornik trudov. Petrozavodsk: Novaya nauka Publ., 2023. P. 105-110.
  20. Koval' N.A. Comparative analysis neural network architectures in the task of detecting and distinguishing target signals and diverting interference. Trudy MAI. 2024. No. 134. (In Russ.). URL: https://trudymai.ru/eng/published.php?ID=17847


Download

mai.ru — informational site MAI

Copyright © 2000-2025 by MAI

 Вход