08. Tarasov Е.М., Andronchev I.K., Bulatov A.A., et al. Development of a Trainable Classifier of State of Rail Lines with Multiple Patterns of Image Recognition

UDK 004.93-047.58

DOI: 10.15507/2658-4123.030.202004.659-682

Development of a Trainable Classifier of State of Rail Lines with Multiple Patterns of Image Recognition

Evgeniy M. Tarasov
Head of Electrical Engineering Chair, Samara State Transport University (2V Svoboda St., Samara 443066, Russian Federation), D.Sc. (Engineering), Professor, Researcher ID: C-2505-2018, ORCID: https://orcid.org/0000-0003-2717-7343, This email address is being protected from spambots. You need JavaScript enabled to view it.

Ivan K. Andronchev
Rector of Samara State Transport University (2V Svoboda St., Samara 443066, Russian Federation), D.Sc. (Engineering), Professor, Researcher ID: AAD-2892-2020, ORCID: https://orcid.org/0000-0002-3964-7050, This email address is being protected from spambots. You need JavaScript enabled to view it.

Andrey A. Bulatov
Vice-Rector of Samara State Transport University (2V Svoboda St., Samara 443066, Russian Federation), Cand.Sc. (Engineering), Associate Professor, Researcher ID: AAC-5665-2020, ORCID: https://orcid.org/0000-0002-1278-2172, This email address is being protected from spambots. You need JavaScript enabled to view it.

Anna E. Tarasova
Postgraduate Student of Electrical Engineering Chair, Samara State Transport University (2V Svoboda St., Samara 443066, Russian Federation), Researcher ID: C-2497-2018, ORCID: https://orcid.org/0000-0001-6907-6036, This email address is being protected from spambots. You need JavaScript enabled to view it.

Introduction. The necessity to classify the state of rail lines affected by significant damaging factors on the sensitive element of the information sensor providing the assurance of classification quality with the required length of the rail lines of the control section forms the task of creating a classifier with extended functionality. Extending the functionality is possible using multidimensional state images with a set of informative features and training procedures for classification models. Using the classical classification principle with a single model leads to an excessive complication of the classification algorithm with low accuracy due to inaccurate solution of the system of conditional equations with multidimensional approximation by Hermite polynomials.
Materials and Methods. The principles of reducing the dimension of the features space, various procedures for trainable classifier of state of rail lines with multidimensional patterns, the selection of decisive classification rules with a hierarchical grouping of classes, and the formation of a set of models of varying degrees of complexity trained to solve an incompatible system of equations are considered to solve the problem. There were obtained various degrees of complexity used in the adaptive algorithm for classifying the rail lines states using Hermite polynomials as models.
Results. The article presents the results of developing 57 classifier models using Hermite polynomials with features of 2, 3, 4, 5, 6 arguments. As an example, the procedure of developing models with 2–6 features is shown. The research results showed that with an increase in the number of features, the quality of classification improves, as when dividing the state space into several classes.
Discussion and Conclusion. The results of the studies confirm the feasibility of the principle of classification of rail line states by a set of classification models, and an algorithm of recursively increasing the classification complexity using a model of increased complexity. The criterion for presenting a new, more complex model is the mismatch between the results of the class calculation by the i-th model and the real class in which the rail line is located at the moment in time.

Keywords: informative features, classifier training, classifier models, classification quality, image

For citation: Tarasov Е.М., Andronchev I.K., Bulatov A.A., et al. Development of a Trainable Classifier of State of Rail Lines with Multiple Patterns of Image Recognition. Inzhenerernyye tekhnologii i sistemy = Engineering Technologies and Systems. 2020; 30(4):659-682. DOI: https://doi.org/10.15507/2658-4123.030.202004.659-682

Contribution of the authors: E. M. Tarasov – formulation of the problem, consulting on theoretical part, analysis of research results, development of mathematical models; I. K. Andronchev – development of an algorithm for classification of rail line states with multiple models, work with literature; A. A. Bulatov – development of algorithms for modeling the recognition of railway line states on computer, analysis of literary sources; A. E. Tarasova – processing of research results, conducting research using the Mathcad, formatting the article.

All authors have read and approved the final manuscript.

Received 14.05.2020; revised 10.07.2020; published online 30.12.2020

REFERENCES

1. Tarasova A.Ye. [Research of Potential Possibilities of Rail State Classifiers]. Nauka i obrazovanie transportu = Science and Education of Transport. 2018; (1):223-226. (In Russ.)

2. Kravtsov Yu.A., Antonov A.A., Bakin M.E. Methods of Interference Immunity Checking of Tone Track Circuits Receivers. Mir transporta = World of Transport and Transportation. 2015; 13(6):102-109. Available at: https://mirtr.elpub.ru/jour/article/view/554 (accessed 03.11.2020). (In Russ.)

3. Mudraya N.G. Automatic Control System of Rail Circuits Rolling on the Basis of Application of Complex Signals. Sovremennye tehnologii. Systemnyi analyz. Modelirovanie = Modern Technologies. System Analysis. Modeling. 2018; 57(1):124-129. Available at: https://clck.ru/RjKTm (accessed 03.11.2020). (In Russ.)

4. Bestemyanov P.F. Energy-Efficient Algorithms for Assessment of the Rail-Circuit Operation. Russian Electrical Engineering. 2017; 88(9):557-562. (In Eng.) DOI: https://doi.org/10.3103/S106837121709005X

5. Tarasov Ye.M. [The Principle of Building a Classifier of Rail Line States by a Set of Informative Features]. Izvestiya Transsiba = News of Transsib. 2012; (1):112-118. Available at: http://izvestia-transsiba.ru/images/journal_pdf/2012-1(9).pdf (accessed 03.11.2020). (In Russ.)

6. Zheleznov D.V, Tarasov E.M., Isaycheva A.G., et al. Development of the Learning Classifier of Rail Lines States with Multivariate Informative Features. Trudy SPIIRAN = SPIIRAS Proceedings. 2017; (1):32-54. (In Russ.) DOI: https://doi.org/10.15622/SP.50.2

7. Rosen Ch.A. Pattern Classification by Adaptive Machines. Science. 1967; 156(3771):38-44. (In Eng.) DOI: https://doi.org/10.1126/science.156.3771.38

8. Vapnik V.N. An Overview of Statistical Learning Theory. IEEE Trans. Neural Networks. 1999; 10(5):988-999. Available at: https://www.math.arizona.edu/~hzhang/math574m/Read/vapnik.pdf (accessed 03.11.2020). (In Eng.)

9. Popov A.A., Boboev Sh.A. Obtaining Sparse Solutions by LS SVM Method through Sample Construction by Optimal Experiment Design Method and Model Quality Criteria. Vestnik Irkutskogo gosudarstvennogo tekhnicheskogo universiteta = Proceedings of Irkutsk State Technical University. 2018; 22(1):100-117. (In Russ.) DOI: https://doi.org/10.21285/1814-3520-2018-1-100-117

10. Lin Ch.-T., Yeh Ch.-M., Liang Sh.-F., et al. Support-Vector-Based Fuzzy Neural Network for Pattern Classification. IEEE Transactions on Fuzzy Systems. 2006; 14(1):31-41. Available at: https://ir.nctu.edu.tw/bitstream/11536/12680/1/000235378000003.pdf (accessed 03.11.2020). (In Eng.)

11. Tarasov Е.М., Gerus V.L., Tarasova А.Е. Study of Informative Value of Features in Rail Condition Monitoring. Vestnik Mordovskogo universiteta = Mordovia University Bulletin. 2018; 28(2):191-206. (In Eng.) DOI: https://doi.org/10.15507/0236-2910.028.201802.191-206

12. Ailem M., Role F., Nadif M. Model-Based Co-Clustering for the Effective Handling of Sparse Data. Pattern Recognition. 2017; 72:108-122. (In Eng.) DOI: https://doi.org/10.1016/j.patcog.2017.06.005

13. Kulkarni K., Turaga P., Srivastava A., et al. Pattern Recognition. New York: Wiley; 2019. (In Eng.) DOI: https://doi.org/10.1002/047134608x.w5513.pub2

14. Rakcheeva T. Focal Model in the Pattern Recognition Problem. In: Advances in Artificial Systems for Medicine and Education II. AIMEE 2018. 2018. Pp. 127-138. (In Eng.) DOI: https://doi.org/10.1007/978-3-030-12082-5_12

15. Nedelko V.M. Some Aspects of Estimating a Quality of Decision Functions Construction Methods. Vestnik Tomskogo gosudarstvennogo universiteta. Upravlenie, vychislitelnaya tekhnika i informatika = Tomsk State University Journal of Control and Computer Science. 2013; (3):123-132. Available at: http://journals.tsu.ru/informatics/&journal_page=archive&id=922&article_id=2381 (accessed 03.11.2020). (In Russ.)

16. Hamed H.A., Elnaz J.H. Pattern Classification. In: Guide to Convolutional Neural Networks. 2017. Pp. 15-83. Available at: https://link.springer.com/chapter/10.1007%2F978-3-319-57550-6_2 (accessed 03.11.2020). (In Eng.)

17. Tarasov E.M., Teplyakov V.B., Gumennikov V.B., et al. On Ensuring Invariance in Problems of Control of Rail-Line Conduction. Russian Electrical Engineering. 2017; 88(3):105-108. (In Eng.) DOI: https://doi.org/10.3103/S1068371217030166

18. Prisukhina I.V., Borisenko D.V. [Machine Classification of the Electric Rail Circuit Operation Mode on the Basis of Logistic Regression]. Omskiy nauchnyy vestnik = Omsk Scientific Bulletin. 2018; (6):126-130. (In Russ.) DOI: https://doi.org/10.25206/1813-8225-2018-162-126-130

19. Amari S. A Theory of Adaptive Pattern Classifiers. IEEE Transactions on Electronic Computers. 1967; 16(3):299-307. (In Eng.) DOI: https://doi.org/10.1109/PGEC.1967.264666

20. Borisenko D.V., Prisukhina I.V., Lunev S.A. [Machine Classification of the Electric Rail Circuit Operation Mode on the Basis of Logistic Regression]. Omskiy nauchnyy vestnik = Omsk Scientific Bulletin. 2018; (4):67-72. (In Russ.) DOI: https://doi.org/10.25206/1813-8225-2018-160-67-72

21. Dokukin A.A., Ryazanov V.V., Shut O.V. Multilevel Models for Solution of Multiclass Recognition Problems. Pattern Recognition and Image Analysis. 2016; 26(3):461-473. (In Russ.) DOI: https://doi.org/10.1134/S1054661816030044

22. Geva S., Sitte J. Adaptive Nearest Neighbor Pattern Classification. IEEE Transactions on Neural Networks. 1991; 2(2):318-322. (In Russ.) DOI: https://doi.org/10.1109/72.80344

23. Prisukhina I.V., Borisenko D.V. [Improvement of Algorithms of Machine Classification of Rail Electrical Systems States as a Part of Automatic Locomotive Signaling]. Omskiy nauchnyy vestnik = Omsk Scientific Bulletin. 2019; (6):63-69. (In Russ.) DOI: https://doi.org/10.25206/1813-8225-2019-168-63-69

24. Zhukov D.A., Kliachkin V.N. The Effect of the Control Sample Volume on the Quality of Diagnostics of the Technical Object State. Avtomatizatsiya protsessov upravleniya = Automation of Control Processes. 2018; (2):90-95. Available at: http://apu.npomars.com/images/pdf/52_11.pdf (accessed 03.11.2020). (In Russ.)

25. Zheleznov D.V., Kotenko A.G., Mikheeva T.I., et al. Development of Invariant Trainable Identifier of Train Coordinate. Vestnik SamGUPS = Bulletin of Samara State University of Railway Transport. 2017; (2):100-112. Available at: https://www.samgups.ru/upload/iblock/414/414b71f869a65cf53ffca34b7e3aa9d1.pdf (accessed 03.11.2020). (In Russ.)

26. Tarasov E.M., Zheleznov D.V. Development of an Intelligent System of Determinating the Coordinates and the Speed of the Train. Transport and Telecommunication Journal. 2016; 17(2):138-143. (In Eng.) DOI: https://doi.org/10.1515/ttj-2016-0013

27. Kravtsov Yu.A., Antonov A.A., Bakin M.E. Methods of Interference Immunity Checking of Tone Track Circuits Receivers. Mir transporta = World of Transport and Transportation. 2015; 13(6):102-109. Available at: https://mirtr.elpub.ru/jour/article/view/554 (accessed 03.11.2020). (In Russ.)

28. Myasnikov E.V. Analysis of Approaches to Feature Space Partitioning for Nonlinear Dimensionality Reduction. Pattern Recognition and Image Analysis. 2016; 26(3):474-482. (In Eng.) DOI: https://doi.org/10.1134/S1054661816030147

29. Liabakh N.N., Umrikhin N.G. Classification Process Automation of Intelligent Transport Complexes on the Basis of Expert Systems. Vestnik Rostovskogo gosudarstvennogo universiteta putey soobshcheniya = Rostov State University of Railway Transport Bulletin. 2013; (2):98-103. Available at: http://vestnik.rgups.ru/?p=22 (accessed 03.11.2020). (In Russ.)

30. Golik F.V. Approximation of the Empirical Probability Distributions by Bernstein Polynomials. Zhurnal radioelektroniki = Journal of Radio Electronics. 2018; (7). (In Eng.) DOI: https://doi.org/10.30898/1684-1719.2018.7.5

31. Kazanskiy N.L., Popov S.B. Integrated Design Technology for Computer Vision Systems in Railway Transportation. Pattern Recognition and Image Analysis. 2015; 25(2):215-219. (In Eng.) DOI: https://doi.org/10.1134/S1054661815020133

32. Klyachkin V.N., Kuvayskova Yu.E., Zhukov D.А. The Use of Aggregate Classifiers in Technical Diagnostics, Based on Machine Learning. Information Technology and Nanotechnology. 2017; 1903:32-35. Available at: http://ceur-ws.org/Vol-1903/paper7.pdf (accessed 03.11.2020). (In Russ.)

33. Tarasov Е.М., Zheleznov D.V., Vasin N.N., et al. Providing Invariance to Disturbing Effects in Rail Lines. Inzhenernyye tekhnologii i sistemy = Engineering Technologies and Systems. 2019; 29(2):152-168. (In Russ.) DOI: https://doi.org/10.15507/2658-4123.029.201902.152-168

This work is licensed under a Creative Commons Attribution 4.0 License.