Method of low-rank online correction of vehicle tire-force estimation in an information-measuring system based on physical constraints

The paper considers online correction of vehicle tire-force estimates in a measurement information system under changes in tire parameters and road adhesion conditions. Direct measurement of longitudinal and lateral forces under operating conditions is difficult; therefore, the initial estimate is formed using a physical baseline model and an offline-trained residual Transformer model. The study focuses not on constructing a complete estimation scheme from scratch, but on low-rank correction of the output layer of a frozen estimation model. The proposed method keeps the parameters of the main model unchanged and updates only the correction block. The main scheme does not use tire-force pseudo-labels; the weighted least-squares pseudo-label variant is considered only as a related methodological technique. The objective function includes lateral-dynamics consistency, yaw-moment consistency, a friction-ellipse constraint and anchor regularization. To select informative intervals, an update-triggering mechanism based on output-estimate jumps and dynamic excitation is used. The computational experiment was performed for five operational-shift scenarios, with ten independent runs performed for each scenario. The trigger rate ranges from 1.88 to 2.27 %, and the update-window coverage ranges from 11.2 to 14.4 %. True tire-force values were not used in the update-release decision and were employed only for subsequent offline evaluation; across the examined scenarios, the reduction in root-mean-square error ranged from zero to 6.23 %. The results show that the effectiveness of online correction depends on the type of operational shift and on the update settings.

1. Pacejka H.B. Tire and Vehicle Dynamics. Oxford: Butterworth-Heinemann; 2012. 672 p.

2. Singh K.Bh., Arat M.A., Taheri S. Literature review and fundamental approaches for vehicle and tire state estimation. Vehicle System Dynamics. 2018;57(11):1643–1665. https://doi.org/10.1080/00423114.2018.1544373

3. Ding X., Wang Zh., Zhang L., Liu J. A Comprehensive Vehicle Stability Assessment System Based on Enabling Tire Force Estimation. IEEE Transactions on Vehicular Technology. 2022;71(11):11571–11588. https://doi.org/10.1109/TVT.2022.3193139

4. Denisenko V.V. Computer control of technological process, experiment, and equipment. Moscow: Goryachaya liniya-Telekom; 2009. 608 p. (In Russ.).

5. Nguyen A.-T., Frezzatto L., Guerra Th.-M., Delprat S. Cost-Effective Estimation of Vehicle Lateral Tire-Road Forces and Sideslip Angle via Nonlinear Sampled-Data Observers: Theory and Experiments. IEEE/ASME Transactions on Mechatronics. 2024;29(6):4606–4617. https://doi.org/10.1109/TMECH.2024.3382777

6. Xu N., Askari H., Huang Y., Zhou J., Khajepour A. Tire Force Estimation in Intelligent Tires Using Machine Learning. IEEE Transactions on Intelligent Transportation Systems. 2022;23(4):3565–3574. https://doi.org/10.1109/TITS.2020.3038155

7. Hongxun F., Zhang J., Luo X., et al. An intelligent tire force estimation correction method based on wheel spoke strain. Sensors and Actuators A: Physical. 2024;380(06):116023. https://doi.org/10.1016/j.sna.2024.116023

8. Acosta M., Kanarachos S., Fitzpatrick M.E. A Virtual Sensor for Integral Tire Force Estimation using Tire Model-less Approaches and Adaptive Unscented Kalman Filter. In: Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics: Volume 1, 26–28 July 2017, Madrid, Spain. SciTePress; 2017. P. 386–397. https://doi.org/10.5220/0006394103860397

9. Løwenstein K.F., Bernardini D., Fagiano L., Bemporad A. Physics-informed online learning of gray-box models by moving horizon estimation. European Journal of Control. 2023;74(4):100861. https://doi.org/10.1016/j.ejcon.2023.100861

10. Raissi M., Perdikaris P., Karniadakis G.E. Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics. 2019;378:686–707. https://doi.org/10.1016/j.jcp.2018.10.045

11. Karniadakis G.E., Kevrekidis I.G., Lu L., Perdikaris P., Wang S., Yang L. Physics-informed machine learning. Nature Reviews Physics. 2021;3(6):422–440. https://doi.org/10.1038/s42254-021-00314-5

12. Hu E.J., Shen Y., Wallis Ph., et al. LoRA: Low-Rank Adaptation of Large Language Models. In: The Tenth International Conference on Learning Representations, 25–29 April 2022, Virtual Event. OpenReview; 2022. URL: https://openreview.net/forum?id=nZeVKeeFYf9

13. Houlsby N., Giurgiu A., Jastrzebski S., et al. Parameter-Efficient Transfer Learning for NLP. In: Proceedings of the 36th International Conference on Machine Learning: Volume 97, 09–15 June 2019, Long Beach, CA, USA. PMLR; 2019. P. 2790–2799. URL: https://proceedings.mlr.press/v97/houlsby19a.html

14. Chernyshev N.N., Alfara A.Y.A., Nizhenec T.V. Automatic control system for vehicle platoon movement. Modeling, Optimization and Information Technology. 2025;13(4). (In Russ.). https://doi.org/10.26102/2310-6018/2025.51.4.016

15. Chugunov M.V., Polunina I.N., Ovchinnikov A.Y., Garikov A.S., Osipov A.V. Technological aspects unmanned electric vehicle design based on the Bigo.Land set and ArduPilot. Modeling, Optimization and Information Technology. 2025;13(3). (In Russ.). https://doi.org/10.26102/2310-6018/2025.50.3.018