Russian Federation
Stability of railway automation systems, utilizing automatic locomotive signaling as a mean for interval regulation of train traffic, can be risen by the means of application of machinal training algorithms. Known algorithm for code signal identification of locomotive signaling has a row of drawbacks which are conditioned by the cyclic character of being identified signals and the necessity of cyclic synchronization before the identification procedure This complicates technical realization essentially as well as does requirements to computational resources of known algorithm. This article offers improved algorithm for identification of code signals transferred via rail chains in the process of train traffic which doesn’t demand cycle synchronization. For to solve cycle synchronization problem it’s been offered to accomplish code signal identification by Fourier amplitude spectrum of its preselected envelope. Such settlement has become possible on account of invariance of Fourier amplitude spectrum of code signal envelope relatively code signal shift in the time domain on random number of digital readings. As a machinal training method, the article considers artificial neural network with simplified, compared with the previous analogues, architecture. While identification, three-layer neural network with full connections between layers is engaged. Also, the work investigates limitations of suggested method of machinal identification which are conditioned by refuse from phase spectrum of envelope of being processed code signal. It’s been demonstrated in the article the act of improved method of machinal identification on real signals fixed in being exploited rail lines. The obtained results testify the adequacy of the suggested methodology for the identification of signals of automatic locomotive signaling of systems for train traffic interval regulation.
machinal training, Fourie transform, automatic locomotive signaling, correlation analysis, impulse interferences, rail line, rail chain
1. Rozenberg E.N., Abramov A.A., Batraev V.V. Interval'noe regulirovanie dvizheniya poezdov // Zheleznodorozhnyy transport. - 2017. - №9. - S. 19-24.
2. Gumarov A.G. Sposoby modulyacii i ih primenenie na zheleznodorozhnom transporte. // Teoriya i praktika sovremennoy nauki - 2020. - №5. - S. 143-154.
3. Pul'tyakov A.V., Skorobogatov M.E. Sistemnyy analiz ustoychivosti raboty sistem avtomaticheskoy lokomotivnoy signalizacii // Sovremennye tehnologii. Sistemnyy analiz. Modelirovanie. - 2018. - №1. - S. 79-89.
4. Yusupov R.R. O korrelyacionnoy deshifracii kodovyh kombinaciy signala ALSN. // Nauka i obrazovanie transportu. - 2020. - №1. - S. 143-154.
5. Pat. 165420 (RF). Priemnoe ustroystvo avtomaticheskoy lokomotivnoy signalizacii / V.B. Leushin, R.R. Yusupov, K.E. Blachev, 2016
6. Pat. 2314223 (RF). Sposob deshifrirovaniya signalov avtomaticheskoy lokomotivnoy signalizacii i ustroystvo dlya ego realizacii / M.D. Rabinovich, B.D. Nikiforov, A.N. Sokolov, A.E. Pravdolyubov, Yu.A. Kravcov, 2005
7. Pat. 2618616 (RF). Ustroystvo podavleniya impul'snyh pomeh na vhode lokomotivnogo priemnika ALS / I.A. Argunov, N.Yu. Vihrova, E.V. Gorenbeyn, E.N. Rozenberg, 2017
8. Pat. 2517631 (RF). Priemnoe ustroystvo avtomaticheskoy lokomotivnoy signalizacii / V.S. Lochehin, 2014
9. Prisuhina I.V., Borisenko D.V. Mashinnaya klassifikaciya signalov chislovogo koda v elektrotehnicheskih sistemah lokomotivnoy signalizacii // Omskiy nauchnyy vestnik. - 2019. - № 4 (166). - S. 39-47.
10. Dem'yanov V.V., Pul'tyakov A.V., Skorobogatov M.E., Alekseenko V.A. Metodika opredeleniya porogovogo znacheniya otnosheniya signal/pomeha dlya sistem avtomaticheskoy lokomotivnoy signalizacii // Avtomatika na transporte - 2020. - №2. - S. 149-164.
11. Prisuhina, I. V., Borisenko D.V., Lunev S.A. Imitacionnaya model' elektricheskogo kodovogo signala v rossiyskih sistemah interval'nogo regulirovaniya dvizheniya poezdov na osnove rel'sovyh cepey // Trudy SPIIRAN. - 2019. - T18. - №5. - S. 1212-1238.
12. Smith S.W. Digital signal processing: a practical guide for engineers and scientists. London, England: Newnes, 2013. 650 p.
13. Oppenheim A., Schafer R. Discrete-time signal processing. London, England: Pearson, 2009. 1144 p.
14. A library for computing the discrete Fourier transform. Available at: http://www.fftw.org (accessed: 06.09.202021) (In English)
15. Goodfellow I., Bengio Y., Courville A. Deep Learning. Cambridge, Massachusetts, USA: MIT Press, 2016. 775 p.
16. Hastie T. The elements of statistical learning. NY: Springer, 2009. 745 p
17. Bishop C. M. Pattern recognition and machine learning. New York, USA: Springer, 2011. 738 p.
18. Mitchel T. M. Machine learning. NY: McGraw-Hill Science, 1997. 432 p.
19. Tyurin S.V. Rezervirovannyy mazhoritarnyy element // Vestnik Permskogo nacional'nogo issledovatel'skogo universiteta. Elektrotehnika, informacionnye tehnologii, sistemy upravleniya - 2018. - №27. - S. 139-152.
