M.A. Petrov1, P.E. Chibizov2, E.V. Melikhova3, A.M. Akkanen4, M.P. Nechaev5, A.N. Briko6

1, 2, 4–6 Bauman Moscow State Technical University (Moscow, Russia)
3 Sirius University of Science and Technology (Federal territory «Sirius», Russia)
1 petrovma@student.bmstu.ru

Modern bionic prostheses can use biological signals of various physical natures. Most often, due to its extensive knowledge and scope of recording, the electromyography signal is used as a control signal. However, its use may be limited by physical factors and physiological characteristics of each person. To overcome these problems, it is necessary to research and develop more accurate and adaptive EMG signal processing algorithms that can compensate for individual user characteristics and external interference.

Consider and analyze approaches of implementing bionic control based on electromyography using classifier models based on machine learning and neural networks and increasing classification accuracy by varying the set of signal features used

The random forest classifier turned out to be the most accurate one in the time and frequency domains – the accuracy was 0.89 and 0.88, respectively. For most classifiers, except for the convolutional neural network, the accuracy in the time domain is higher than in the frequency domain. A set of features providing the highest accuracy is: Wilson amplitude, number of zero crossings, maximum amplitude.

Defined sets of features and preprocessing methods make it possible to achieve the highest classification accuracy with less computation. The use of more accurate classifiers improves the classification accuracy of the control systems in bionic devices.

Petrov M.A., Chibizov P.E., Melikhova E.V., Akkanen A.M., Nechaev M.P., Briko A.N. Algorithmic strategies for processing EMG signals to provide bionic control of upper limb prostheses. Biomedicine Radioengineering. 2024. V. 27. № 6. P. 30–39. DOI: https:// doi.org/10.18127/j15604136-202406-03 (In Russian)

1. Briko A., Kapravchuk V., Dyachencova S. Biotechnical control system based on electromyogram, electroimpedance and myotonogram signals. 2021 International Conference on Engineering and Emerging Technologies (ICEET). IEEE. 2021. P. 1–4. DOI: 10.1109/ ICEET53442.2021.9659559
2. Farfán F.D., Politti J.C., Felice C.J. Evaluation of EMG processing techniques using information theory. Biomedical engineering online. 2010. V. 9. P. 1–18. DOI: 10.1186/1475-925X-9-72
3. Rubinov N.S., Rudakov I.V., Stroganov I.V. Development of Specialized Software for Biomedical Research. 2020 IEEE Confe¬rence of Russian Young Researchers in Electrical and Electronic Engineering (EIConRus). IEEE. 2020. P. 486–488. DOI: 10.1109/EIConRus49466.2020.9039318
4. Park S.H, Lee S.P. EMG Pattern Recognition Based on Artificial Intelligence Techniques. IEEE Transactions on Rehabilitation Engi-neering. 1998. V. 6. № 4. P. 400–405. DOI: 10.1109/86.736154
5. Hermens H.J., Freriks B., Disselhorst-Klug C., Rau G. Development of recommendations for sEMG sensors and sensor placement procedures. Journal of electromyography and Kinesiology. 2000. V. 5. № 10. P. 361–374. DOI: 10.1016/S1050-6411(00)00027-4
6. Kotov-Smolenskiy A.M., Khizhnikova A.E., Klochkov A.S., Suponeva N.A., Piradov M.A. Surface EMG: applicability in the motion analysis and opportunities for practical rehabilitation. Human Physiology. 2021. V. 47. № 2. P. 237–247. DOI: 10.1134/S0362119721020043
7. SENIAM project. Hermens H.J., Freriks B. URL: http://www.seniam.org (data obrashcheniya: 05.04.2024).
8. Aigner R., Wigdor D., Benko H., Haller M. Understanding mid-air hand gestures: A study of human preferences in usage of gesture types for hci. Microsoft Research TechReport MSR-TR-2012-111. V. 2. № MSR-TR-2012-111. P. 30.
9. Rota S., Rogowski I., Champely S., Hautier C. Reliability of EMG normalisation methods for upper-limb muscles. Journal of sports sciences. 2013. № 31. P. 1696–1704.
10. Parajuli N., Sreenivasan N., Bifulco P., Cesarelli M. Real-time EMG based pattern recognition control for hand prostheses: A review on existing methods, challenges and future implementation. Sensors. 2019. V. 19. № 20. P. 4596. DOI: 10.3390/s19204596
11. Gopal P., Gesta A., Mohebbi A. A systematic study on electromyography-based hand gesture recognition for assistive robots using deep learning and machine learning models. Sensors. 2022. V. 22. № 10. P. 3650. DOI: 10.3390/s22103650
12. Li W., Shi P., Yu H. Gesture recognition using surface electromyography and deep learning for prostheses hand: state-of-the-art, chal-lenges, and future. Frontiers in neuroscience. 2021. V. 15. P. 621885. DOI: 10.3389/fnins.2021.621885
13. Rajapriya R., Rajeswari K., Thiruvengadam S.J. Deep learning and machine learning techniques to improve hand movement classifica-tion in myoelectric control system. Biocybernetics and Biomedical Engineering. 2021. V. 41. № 2. P. 554–571. DOI: 10.1016/j.bbe. 2021.03.006
14. Kobelev A.V., Shchukin S.I. Antropomorfnoe upravlenie protezom predplech'ya na osnove elektroimpedansnoj miografii. Fizicheskie osnovy priborostroeniya. 2019. T. 8. №. 4. S. 62–68. DOI: 10.25210/jfop-1904-062068
15. Altmann A., Toloşi L., Sander O., Lengauer T. Permutation importance: a corrected feature importance measure. Bioinformatics. 2010. V. 26. № 10. P. 1340–1347. DOI: 10.1093/bioinformatics/btq134