Modern Transportation Systems and TechnologiesModern Transportation Systems and TechnologiesИнновационные транспортные системы и технологии2782-3733Eco-Vector65980910.17816/transsyst659809Original studiesОригинальные статьиResearch ArticleElectronic differential system based on neural networks for electric vehicles: development, adaptation and prospects of applicationСистема электронного дифференциала на основе нейронных сетей для электромобилей: развитие, адаптация и перспективы примененияhttps://orcid.org/0000-0001-7282-84701956-3662LisovAndrey A.ЛисовАндрей Анатольевич
Russian Federation

postgraduate student

аспирант

lisov.andrey2013@yandex.ruhttps://orcid.org/0000-0002-1292-39752893-8730VozmilovAlexander G.ВозмиловАлександр Григорьевич
Russian Federation

Professor, Doctor of Technical Sciences

профессор, д.т.н

vozmiag@rambler.ruSouth Ural State UniversityЮжно-Уральский государственный университет1503202511124422102202524022025Copyright ©; 2025, Lisov A.A., Vozmilov A.G.Copyright ©; 2025, Лисов А.А., Возмилов А.Г.2025Lisov A.A., Vozmilov A.G.Лисов А.А., Возмилов А.Г.https://creativecommons.org/licenses/by/4.0https://transsyst.ru/transj/article/view/659809

Aim. The analysis of possibilities and prospects of development of an electronic differential system for electric vehicles based on artificial neural networks.

Materials and Methods. We discuss the key advantages of the proposed system, such as its customization capability to various vehicle designs, integration of additional sensors, support for self-driving mode and the ability to interact with the ABS system.

Results. We considered the ways to improve the model, including the introduction of self-learning algorithms, optimization of inverter circuits for controlling multiple motors, and implementation of all-wheel drive configurations. In addition, we discuss the customization of the electronic differential system for operation on low-power devices using quantization, pruning and architecture simplification methods.

Conclusion. The proposed approaches and algorithms have the potential for widespread deployment in the electric vehicle industry, opening new vistas for development of intelligent vehicle control systems.

Цель. Анализ возможностей и перспектив развития системы электронного дифференциала (СЭД) для электромобилей, основанной на использовании искусственных нейронных сетей (ИНС).

Материалы и методы. Обсуждаются ключевые преимущества предложенной системы, такие как ее адаптивность к различным конструктивным конфигурациям автомобиля, интеграция дополнительных датчиков, поддержка беспилотного режима и возможность взаимодействия с системой ABS.

Результаты. Рассмотрены пути совершенствования модели, включая внедрение самообучающихся алгоритмов, оптимизацию инверторных схем для управления несколькими двигателями и реализацию полноприводных конфигураций. Также обсуждается адаптация СЭД для работы на маломощных устройствах с использованием методов квантизации, прореживания (pruning) и упрощения архитектуры.

Заключение. Предложенные подходы и алгоритмы имеют потенциал для широкого внедрения в индустрии электромобилей, открывая новые возможности для развития интеллектуальных систем управления транспортными средствами.

electronic differentialelectric vehicleneural networksself-learninginverterself-drivingoptimizationABS integrationэлектронный дифференциалэлектромобильнейронные сетисамообучениеинверторавтономное вождениеоптимизацияинтеграция ABSLisov AA. Development of an electronic differential system for electric vehicles based on deep neural network. Modern Transportation Systems and Technologies. 2024;10(3):351–367. (in Russ.) doi: 10.17816/transsyst634127Лисов А.А. Разработка системы электронного дифференциала для электромобилей на основе нейросети глубокого обучения // Инновационные транспортные системы и технологии. 2024 Т. 10. № 3. С. 351–367. doi: 10.17816/transsyst634127Lisov AA, Panishev SA, Gundarev KA. Development of an Electronic Differential System Based on Artificial Neural Networks for Electric Transport. In: 2024 International Ural Conference on Electrical Power Engineering (UralCon). 2024;192–196. doi: 10.1109/UralCon62137.2024.10718938Lisov A.A., Panishev S.A., Gundarev K.A. Development of an Electronic Differential System Based on Artificial Neural Networks for Electric Transport. In: 2024 International Ural Conference on Electrical Power Engineering (UralCon). 2024 P. 192–196. doi: 10.1109/UralCon62137.2024.10718938Yildirim M, Kurum H. Electronic differential system for an electric vehicle with four in-wheel PMSM. In: Proceedings of the 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring). 2020;1–5. doi: 10.1109/VTC2020-Spring48590.2020.9129139Yildirim M., Kurum H. Electronic differential system for an electric vehicle with four in-wheel PMSM // Proceedings of the 2020 IEEE 91st Vehicular Technology Conference (VTC2020-Spring). 2020. P. 1–5. doi: 10.1109/VTC2020-Spring48590.2020.9129139Vozmilov AG, Urmanov VG, Lisov AA, Ilimbetov RYu. Development and simulation of the proportional and integral regulator for the electronic differential of an electric farm tractor. Bulletin of the Bashkir State Agrarian University. 2022;2(62):119–124 (In Russ.). doi: 10.31563/1684-7628-2022-62-2-119-124Возмилов А.Г., Урманов В.Г., Лисов А.А., Илимбетов Р. Ю. Разработка и моделирование пропорционально-интегрального регулятора для электронного дифференциала электротрактора сельскохозяйственного назначения // Вестник Башкирского Государственного Аграрного Университета. 2022. Т. 2, № 62. С. 119–124. doi: 10.31563/1684-7628-2022-62-2-119-124Rigas ES, Ramchurn SD, Bassiliades N. Managing electric vehicles in the smart grid using artificial intelligence: A survey. IEEE Transactions on Intelligent Transportation Systems. 2014;16(4):1619–1635. doi: 10.1109/TITS.2014.2376873Rigas E.S., Ramchurn S.D., Bassiliades N. Managing electric vehicles in the smart grid using artificial intelligence: A survey // IEEE Transactions on Intelligent Transportation Systems. 2014. Vol. 16(4). P. 1619–1635. doi: 10.1109/TITS.2014.2376873Nguyen DH, Widrow B. Neural networks for self-learning control systems. IEEE Control Systems Magazine. 2002;10(3):18–23. DOI: 10.1109/37.55119Nguyen D.H., Widrow B. Neural networks for self-learning control systems // IEEE Control Systems Magazine. 2002. Vol. 10(3). P. 18–23. doi: 10.1109/37.55119Li SE. Deep reinforcement learning. In: Reinforcement Learning for Sequential Decision and Optimal Control. Singapore: Springer Nature Singapore. 2023:365–402. doi: 10.1007/978-981-19-7784-8_10Li S. E. Deep reinforcement learning. In: Reinforcement Learning for Sequential Decision and Optimal Control. Singapore: Springer Nature Singapore. 2023. P. 365–402. doi: 10.1007/978-981-19-7784-8_10Hizume M. Independent vector control of parallel-connected two induction motors by a five-leg inverter. In: Proceedings of the 10th European Conference on Power Electronics and Applications. 2003;778.Hizume M. Independent vector control of parallel-connected two induction motors by a five-leg inverter. In: Proceedings of the 10th European Conference on Power Electronics and Applications. 2003. Vol. 778.Nozawa Y. Independent position control of two permanent magnet synchronous motors with five-leg inverter by the expanded two arm modulation method. JIASC. 2005;79.Nozawa Y. Independent position control of two permanent magnet synchronous motors with five-leg inverter by the expanded two arm modulation method // JIASC. 2005. Vol. 79.Oka K. Improved method of voltage transfer ratio for five-leg inverter. JIASC2005. 2005;75.Oka K. Improved method of voltage transfer ratio for five-leg inverter. // JIASC2005. 2005. Vol. 75.Gulbudak O, Gokdag M. Dual-hysteresis band control of nine-switch inverter to control two induction motors. IEEE Transactions on Energy Conversion. 2021;37(2):788–99. doi: 10.1109/TEC.2021.3131385Gulbudak O, Gokdag M. Dual-hysteresis band control of nine-switch inverter to control two induction motors // IEEE Transactions on Energy Conversion. 2021. Vol. 37(2). P. 788–99. doi: 10.1109/TEC.2021.3131385Krishnakumar V, Kamaraj V, Perianayagam CA. An integrated drive for two PMSMs involved automotive applications and development of current reference expanded two arm modulation technique. Circuits and Systems. 2016;7(08):1794–1799. doi: 10.4236/cs.2016.78155Krishnakumar V, Kamaraj V, Perianayagam C.A. An integrated drive for two PMSMs involved automotive applications and development of current reference expanded two arm modulation technique // Circuits and Systems. 2016. Vol. 7(08). P. 1794–1799. doi: 10.4236/cs.2016.78155Pao YH, Phillips SM. The functional link net and learning optimal control. Neurocomputing. 1995;9:149–164. doi: 10.1016/0925-2312(95)00066-FPao Y.H., Phillips S.M. The functional link net and learning optimal control // Neurocomputing. 1995. Vol. 9. P.149–164. doi: 10.1016/0925-2312(95)00066-FRadunovic B, Le Boudec JY. A unified framework for max-min and min-max fairness with applications. IEEE/ACM Transactions on networking. 2007;15(5):1073–1083. doi: 10.1109/TNET.2007.896231Radunovic B., Le Boudec J.Y. A unified framework for max-min and min-max fairness with applications // IEEE/ACM Transactions on networking. 2007. Vol. 15(5). P. 1073–1083. doi: 10.1109/TNET.2007.896231Razaviyayn M, Huang T, Lu S, Nouiehed M, Sanjabi M, Hong M. Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances. IEEE Signal Processing Magazine. 2020;37(5):55–66. doi: 10.1109/MSP.2020.3003851Razaviyayn M., Huang T., Lu S., et al. Nonconvex min-max optimization: Applications, challenges, and recent theoretical advances // IEEE Signal Processing Magazine. 2020. Vol. 37(5). P. 55–66. doi: 10.1109/MSP.2020.3003851Kim N, Bae J, Kim C, et al. Object distance estimation using a single image taken from a moving rolling shutter camera. Sensors. 2020;20(14):3860–3870. doi: 10.3390/s20143860Kim N., Bae J., Kim C., et al. Object distance estimation using a single image taken from a moving rolling shutter camera // Sensors. 2020 Vol. 20(14). P. 3860–3870. doi: 10.3390/s20143860Zhu W, Anderson B, Zhu S, Wang YA. computer vision-based system for stride length estimation using a mobile phone camera. In: 18th International ACM SIGACCESS Conference on Computers and Accessibility. 2016:121–130. doi: 10.1145/2982142.298215Zhu W., Anderson B., Zhu S., Wang YA. computer vision-based system for stride length estimation using a mobile phone camera. In: 18th International ACM SIGACCESS Conference on Computers and Accessibility. 2016. P. 121–130. doi: 10.1145/2982142.298215Gholami A, Kim S, Dong Z, et al. A survey of quantization methods for efficient neural network inference. Low-Power Computer Vision. 2022:291–326.Gholami A., Kim S., Dong Z., Yao Z., Mahoney M.W., Keutzer K. A survey of quantization methods for efficient neural network inference // Low-Power Computer Vision. 2022. P. 291–326.Guo Y. A survey on methods and theories of quantized neural networks. arXiv preprint arXiv:1808.04752. 2018. doi: 10.48550/arXiv.1808.04752Guo Y. A survey on methods and theories of quantized neural networks // arXiv preprint arXiv:1808.04752. 2018. doi: 10.48550/arXiv.1808.04752Blalock D, Gonzalez Ortiz JJ, Frankle J, Guttag J. What is the state of neural network pruning? Proceedings of machine learning and systems. 2020;2:129–146.Blalock D., Gonzalez Ortiz J.J., Frankle J., Guttag J. What is the state of neural network pruning? // Proceedings of machine learning and systems. 2020. Vol. 2. P. 129–146.Cheng H, Zhang M, Shi JQ. A survey on deep neural network pruning: Taxonomy, comparison, analysis, and recommendations. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024. doi: 10.1109/TPAMI.2024.3447085Cheng H., Zhang M., Shi J.Q. A survey on deep neural network pruning: Taxonomy, comparison, analysis, and recommendations // IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024. doi: 10.1109/TPAMI.2024.3447085Cheng Y, Wang D, Zhou P, Zhang T. Model compression and acceleration for deep neural networks: The principles, progress, and challenges. IEEE Signal Processing Magazine. 2018;35(1):126–36. doi: 10.1109/MSP.2017.2765695Cheng Y., Wang D., Zhou P., Zhang T. Model compression and acceleration for deep neural networks: The principles, progress, and challenges // IEEE Signal Processing Magazine. 2018. Vol. 35(1). P. 126–36. doi: 10.1109/MSP.2017.2765695Deng L, Li G, Han S, et al. Model compression and hardware acceleration for neural networks: A comprehensive survey. Proceedings of the IEEE. 2020;108(4):485–532. doi: 10.1109/JPROC.2020.2976475Deng L., Li G., Han S., Shi L., Xie Y. Model compression and hardware acceleration for neural networks: A comprehensive survey // Proceedings of the IEEE. 2020. Vol. 108(4). P. 485–532. doi: 10.1109/JPROC.2020.2976475Burgueno L, Cabot J, Li S, Gérard S. A generic LSTM neural network architecture to infer heterogeneous model transformations. Software and Systems Modeling. 2022;21(1):139–156. doi: 10.1007/s10270-021-00893-yBurgueno L., Cabot J., Li S., Gérard S. A generic LSTM neural network architecture to infer heterogeneous model transformations // Software and Systems Modeling. 2022. Vol. 21(1). P. 139–156. doi: 10.1007/s10270-021-00893-yBurgueno L, Cabot J, Li S, Gérard S. An LSTM-based neural network architecture for model transformations. ACM/IEEE 22nd International Conference on Model Driven Engineering Languages and Systems (MODELS). 2019:294–299. doi: 10.1016/j.cola.2022.101174Burgueno L., Cabot J., Li S., Gérard S. An LSTM-based neural network architecture for model transformations. In: ACM/IEEE 22nd International Conference on Model Driven Engineering Languages and Systems (MODELS). 2019 P. 294–299. doi: 10.1016/j.cola.2022.101174Bulat A, Kossaifi J, Tzimiropoulos G, Pantic M. Matrix and tensor decompositions for training binary neural networks. arXiv preprint arXiv:1904.07852. 2019. doi: 10.48550/arXiv.1904.07852Bulat A., Kossaifi J., Tzimiropoulos G., Pantic M. Matrix and tensor decompositions for training binary neural networks // arXiv preprint arXiv:1904.07852. 2019. doi: 10.48550/arXiv.1904.07852Chen S, Zhou J, Sun W, Huang L. Joint matrix decomposition for deep convolutional neural networks compression. Neurocomputing. 2023;516:11–26. doi: 10.1016/j.neucom.2022.10.021Chen S., Zhou J., Sun W., Huang L. Joint matrix decomposition for deep convolutional neural networks compression // Neurocomputing. 2023. Vol. 516. P. 11–26. doi: 10.1016/j.neucom.2022.10.021Capra M, Bussolino B, Marchisio A, et al. Hardware and software optimizations for accelerating deep neural networks: Survey of current trends, challenges, and the road ahead. IEEE Access. 2020;8:225134–225180. doi: 10.1109/ACCESS.2020.3039858Capra M., Bussolino B., Marchisio A., et al. Hardware and software optimizations for accelerating deep neural networks: Survey of current trends, challenges, and the road ahead // IEEE Access. 2020. Vol. 8. P. 225134–225180. doi: 10.1109/ACCESS.2020.3039858