Machine learning models comparison for vibration-based classification of defects presence in robotics components

Machine learning models comparison for vibration-based classification of defects presence in robotics components

Aleksandr Yu. Tamm
Head of the Research Center, Russian State Scientific Center for Robotics and Technical Cybernetics (RTC), Research Center, 21, Tikhoretsky pr., Saint Petersburg, 194064, Russia, This email address is being protected from spambots. You need JavaScript enabled to view it., ORCID: 0000-0001-5392-1389

Yuri Ya. Boldyrev
Doctor of Engineering Sciences, Professor, Peter the Great St. Petersburg Polytechnic University (SPbPU), 29B, Polytechnicheskaya ul., Saint Petersburg, 195251, Russia, This email address is being protected from spambots. You need JavaScript enabled to view it.

Elizaveta A. Barymova
Mathematician, RTC, 21, Tikhoretsky pr., Saint Petersburg, 194064, Russia, This email address is being protected from spambots. You need JavaScript enabled to view it., ORCID: 0000-0001-6511-2060

UDC identifier: 004.94

EDN: FZQEEA

Abstract. The paper is devoted to the description of the results comparison for different machine learning models determining the presence of defects in the bearing of the wave gearbox in robot using weakly structured data. The comparison was made using a synthetic dataset, which has been obtained with the finite element method. Training was performed on a small number of examples. The quality of classification was evaluated with respect to the learning and classification time. According to the results, the most promising model of machine learning for further use was selected under the limitation on the number of training examples and allowing the classification of the state in the presence of several defects in the structure at the same time.

Key words: machine learning, segment-spike neural networks, diagnostics of complex technical systems, numerical modeling

For citation: Tamm, A.Yu., Boldyrev, Yu.Ya. and Barymova, E.A. (2025), "Machine learning models comparison for vibration-based classification of defects presence in robotics components", Robotics and Technical Cybernetics, vol. 13, no. 3, pp. 191-197, EDN: FZQEEA. (in Russian).

Acknowledgements
The results were obtained within the framework of the state assignment from the Ministry of Science and Higher Education of Russia, «Research on methods for analyzing weakly structured data, knowledge processing, and the creation of cognitive agents based on combined deep neural networks» (Project No. FNRG-2025-0008 1024050200009-5-1.2.1;2.2.2).

References

  1. Zhuravlev, Yu.I., Ryazanov, V.V. and Sen'ko, O.V. (2026), Raspoznavanie. Matematicheskie metody. Programmnaya sistema. Prakticheskie primeneniya [Recognition. Mathematical methods. Software system. Practical applications], Fazis, Moscow, Russia. (in Russian).
  2. Min-Ling Zhang and Zhi-Hua Zhou (2007), “ML-KNN: A lazy learning approach to multi-label learning”, Pattern Recognition, vol. 40, is. 7, 2007, pp. 2038-2048, ISSN 0031-3203, DOI: 10.1016/j.patcog.2006.12.019
  3. Von Winterfeldt, D. and Edwards, W. (1993), Decision analysis and behavioral research, Cambridge University Press, Arch. Decisionanalysis0000vonw.
  4. Breiman, L. (2001), “Random forests”, Machine learning. Kluwer Academic Publishers. Manufactured in The Netherlands, vol. 45, pp. 5-32, available at: https://link.springer.com/article/10.1023/A:1010933404324 (Accessed 7 January 2024)
  5. Tikhonov, A.N. (1965), “On ill-posed problems of linear algebra and a stable method for solving them”, Reports of the USSR Academy of Sciences, vol. 163, no. 3, pp. 591-594. (in Russian).
  6. Ruck, D.W., Rogers, S.K. and Kabrisky, M. (1990), “Feature selection using a multilayer perceptron”, Journal of neural network computing, vol. 2, no. 2, pp. 40-48, available at: https://link.springer.com/chapter/10.1007/11550907_53 (Accessed 7 January 2024)
  7. Scharstein, H. (1979), “Input-output relationship of the leaky-integrator neuron model”, Journal of mathematical biology, vol. 8, no. 4, pp. 403-420, DOI: 10.1007/BF00275835.
  8. Bakhshiev, A.V. and Demcheva, A.A. (2022), “Compartmental spiking neuron model CSNM”, Izvestiya VUZ. Applied Nonlinear Dynamics, vol. 30, iss. 3, pp. 299-310, DOI: 10.18500/0869-6632-2022-30-3-299-310. (in Russian).
  9. Korsakov, A.M, Bakhshiev, A.V., Stankevich, L.A. and Lopota, A.V. (2023), “Neuromorphic tools for solving classification and control problems”, Systems analysis in design and management: collection of scientific papers of the XXVII International scientific and practical conference, St. Petersburg, October 13-14, 2023, St. Petersburg: POLYTECH-PRESS, pp. 220-230, DOI: 10.18720/SPBPU/2/id24-167. (in Russian).
  10. Auge, D., Hille, J., Mueller, E. et al. (2021), “A Survey of Encoding Techniques for Signal Processing in Spiking Neural Networks”, Neural Process Lett 53, pp. 4693–4710, DOI: 10.1007/s11063-021-10562-2
  11. Tamm, A. Yu., Barymova, E.A. and Kuz'min, M.I. (2025), “Study of the application of spiking neural network and finite element method for diagnostics of RTC”, Izvestiya YuFU. Tekhnicheskie nauki, no. 2(244), pp.185-193, DOI: 10.18522/2311-3103-2025-2-185. (in Russian).

Received 03.06.2025
Revised 13.07.2025
Accepted 27.07.2025