E.D. Larionova1, O.V. Druzhinina2

1, 2 FRС «Computer Science and Control» of RAS (Moscow, Russia)
1еva.sh.2363@gmail.com, 2odruzhinina@frccsc.ru

Modern railway infrastructure diagnostic systems generate large volumes of data which require advanced, energy-efficient processing methods. A promising approach involves the use of intelligent techniques for analyzing track geometry car measurements and predicting track state. A key direction in this field is the implementation of neural networks to enhance the accuracy of diagnostics and forecasting in railway traffic control systems. Furthermore, improving the techniques for assessing the technical condition of railway tracks based on the severity of deviations is a relevant task. In practice, situations often arise where the actual state of the rail track does not align with the standard-defined severity of deviations. Therefore, it is advisable to develop new mathematical models capable of performing deviation calculations that account for various combinations of track irregularity parameters. The objectives of this work are to develop an approach to create an intelligent decision support system for assessing the technical condition of a railway track, and to develop an algorithm for analyzing deviations from standards based on a mathematical model and a neural network algorithm for predicting changes in technical state. The diagnostic and predictive modules of the intelligent decision support system for railway track state assessment are described. A mathematical model for analyzing deviations based on input parameters is considered, an algorithm is developed, and a software implementation of this model is proposed. The issue of data preparation in the case of boundary values of parameters for deviations from the standards is considered. To implement the predictive module a machine learning algorithm using a recurrent neural network is developed. The results of computational experiments within the framework of the studied mathematical model with interpretation of qualitative effects are presented. The results can be applied to the development and enhancement of automated control systems for the technical state of transport systems, as well as to the modeling of technical systems using intelligent analysis.

Larionova E.D., Druzhinina O.V. Development of diagnostic and predictive modules of intelligent decision support system for assessing the railway track technical state. Nonlinear World. 2025. V. 23. № 4. P. 36–42. DOI: https:// doi.org/10. 18127/j20700970-202504-04 (In Russian)

1. Gosudarstvennaya programma «Razvitie transportnoj infrastruktury», utverzhdennaya po-stanovleniem Pravitel'stva RF ot 20.12.2017 № 1596 (In Russian).
2. Strategiya nauchno-tekhnologicheskogo razvitiya holdinga «RZHD» na period do 2025 goda i na perspektivu do 2030 goda (belaya kniga), utverzhdennaya rasporyazheniem OAO RZHD ot 17.04.2018 g. N 769/r (In Russian).
3. Ozerov A.V., Kuropteva A.P. Sovremennye sistemy tekhnicheskoj diagnostiki i monitoringa na zheleznodorozhnom transporte. Nauka i tekhnologii zheleznyh dorog. 2024. T. 8. № 1 (29). S. 3–13. (In Russian).
4. Platonov E.N., Prosvirin K.V. Prognozirovanie defektov verhnego stroeniya zheleznodo-rozhnogo puti metodami mashinnogo obucheniya. Vestnik komp'yuternyh i informacionnyh tekh-nologij. 2022. T. 19. № 2. S. 8–18 (In Russian).
5. Korol'kov E.P., Druzhinina O.V., Larionova E.D. Analiticheskoe i chislennoe issledovanie matematicheskoj modeli dlya kontrolya shiriny rel'sovoj kolei pri otstupleniyah ot normativ-nyh parametrov. Transport: nauka, tekhnika, upravlenie. 2023. № 12. S. 32–37 (In Russian).
6. Instrukciya po tekushchemu soderzhaniyu zheleznodorozhnogo puti, utverzhdennaya rasporyazheni-em OAO «RZHD» ot 14 noyabrya 2016 g. № 2288r (v redakcii rasporyazheniya OAO «RZHD» ot 13 de-kabrya 2023 g. № 3167/r) (In Russian).
7. Instrukciya po ocenke sostoyaniya rel'sovoj kolei puteizmeritel'nymi sredstvami i meram po obespecheniyu bezopasnosti dvizheniya poezdov, utverzhdennaya rasporyazheniem OAO «RZHD» ot 28 fevralya 2020 g. № 436/r (v redakcii rasporyazheniya OAO «RZHD» ot 12 dekabrya 2024 g. № 3099/r) (In Russian).
8. Lyudagovskaya M.A., Larionova E.D. Primenenie gibridnyh nejronnyh setej dlya prognozi-rovaniya izmeneniya tekhnicheskogo sostoyaniya zheleznodorozhnogo puti. Nejrokomp'yutery: razra-botka, primenenie. 2025. T. 27. № 2. S. 62–68 (In Russian).
9. Druzhinina O.V., Loktev A.A. Dmitrashko A.V., Sychev P.V. Issledovanie matematicheskoj modeli prognozirovaniya tekhnicheskogo sostoyaniya zheleznodorozhnogo puti. Transport: nauka, tekhnika, upravlenie. 2019. № 4. S. 3–11 (In Russian).
10. Druzhinina O.V., Korepanov E.R., Petrov A.A., Makarenkova I.V., Maksimova V.V. Razra-botka instrumental'no-metodicheskogo obespecheniya dlya ocenki tekhnicheskogo sostoyaniya uzlov transportnyh sistem s primeneniem nejronnyh setej, algoritmov nechetkogo i stohasticheskogo modelirovaniya. Nejrokomp'yutery: razrabotka, primenenie. 2024. T. 26. № 2. S. 13−22 (In Russian).