RiderID: Investigating Cycling as a Behavioral Biometric through Bicycle-Mounted Sensors
Saad, A., Mahdi, M., Keppel, J. & Matviienko, A.
Proceedings of the 24th International Conference on Mobile and Ubiquitous Multimedia, 2025
Summary
Despite their ubiquity, bicycles remain largely unexplored as a space for personalization and cyclist-aware solutions. Since cycling is inherently a behavioral activity, we investigate its potential as a biometric trait for continuous identification and personalization. For this, we mounted multiple sensors on a regular bicycle and conducted an outdoor experiment (N = 16) across two sessions on two cycling tracks: a short, controlled track and a long, in-the-wild setup. This design enabled us to examine how stable cycling patterns are across sessions and how well individuals can be distinguished in controlled versus real-world conditions. Our findings demonstrate the feasibility of leveraging cycling behavior for continuous identification and personalization, achieving identification accuracies of up to 94.3%. We discuss the implications of these results for real-world deployment, as well as the challenges posed by environmental factors, rider variability, and the need for robust and unobtrusive sensing solutions.
Written for this site, the published abstract is at the DOI.
Cite as
Saad, A., Mahdi, M., Keppel, J. & Matviienko, A. (2025). RiderID: Investigating Cycling as a Behavioral Biometric through Bicycle-Mounted Sensors. In Proceedings of the 24th International Conference on Mobile and Ubiquitous Multimedia (pp. 340–350). https://doi.org/10.1145/3771882.3771901
