A Multimodal Fatigue-Driving-Detection Algorithm Using Federated Learning and DS-Evidence Theory
Vehicles · 07.11.2026 · 8 (11) · 293
DOI: 10.3390/vehicles8110293 ↗Kurz gefasst
Fachartikel in der Zeitschrift Vehicles, online erschienen am 07.11.2026. Verfasst von Yuxuan Su und 1 weiteren Autor:innen. Die Arbeit nennt 20 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Fatigue-driving detection is of great significance for reducing traffic accidents and enhancing the active safety level of vehicles. To solve the problems of insufficient single-modal information and data-privacy protection, in this paper, a multimodal fatigue-driving-detection algorithm using federated learning (FL) and Dempster–Shafer (DS)-evidence theory is proposed. Three types of information (facial image, electroencephalogram (EEG) signal, and vehicle behavior) are proposed to compensate for the instability of single-modal detection. Meanwhile, DS-evidence theory is used for fusion, which can improve the detection performance. Additionally, the FL mechanism is introduced to achieve collaborative training among multiple clients without directly sharing raw data. It solves the problem of data-privacy protection. The experiment results show that, compared with the single-modality-based FL method, the accuracy, recall and F1 score increase by 4.5%, 1.50% and 3.7%, respectively. The AUC-ROC has increased from 0.870 to 0.937 and the EER has decreased from 0.200 to 0.135. Compared with the multimodal fusion of the single client method, the accuracy, recall and F1 score have increased by 2.2%, 2.5% and 1.4%. The AUC-ROC has increased from 0.915 to 0.937 and the EER has decreased from 0.155 to 0.135.
Zitieren
APA
Su, Y., & Yan, J. (2026). A Multimodal Fatigue-Driving-Detection Algorithm Using Federated Learning and DS-Evidence Theory. Vehicles, 8(11), 293. https://doi.org/10.3390/vehicles8110293
Vancouver
Su Y, Yan J. A Multimodal Fatigue-Driving-Detection Algorithm Using Federated Learning and DS-Evidence Theory. Vehicles. 2026;8(11):293. doi:10.3390/vehicles8110293
BibTeX
@article{Su2026,
author = {Su, Yuxuan and Yan, Jun},
title = {A Multimodal Fatigue-Driving-Detection Algorithm Using Federated Learning and DS-Evidence Theory},
journal = {Vehicles},
year = {2026},
volume = {8},
number = {11},
pages = {293},
doi = {10.3390/vehicles8110293}
}
EUVIMED