Can vibration patterns identify users? Authentication for smartphone-watch collaboration
Author Identifier (ORCID)
Abstract
With increasing popularity of mobile smart devices user privacy and security are becoming particularly critical. Traditional authentication methods such as passwords and fingerprints, face the risk of forgery attacks and replay attacks, which are difficult to provide anti-imitation identity discrimination. This paper investigates the propagation characteristics of active vibration signals in hand-wrist coordination during natural touch operations by users, and models their response patterns. We found that there are differences in hand-wrist structure and micro-dynamic behavior among different individuals, which exhibits stable and distinguishable dynamic characteristics in vibration response. Based on the observation, we propose a dual-terminal joint identity authentication system (called VIP) for both smartphone and smartwatch. The system first generates vibration by dual-terminal motors and collects response signals using built in accelerometers, gyroscopes, and magnetometers to extract differential response patterns of users in natural touch. Then, we design a discriminative model, TouchFormer, that can extract direction sensitive features in multi-axis vibrations and achieve dynamic alignment and time compensation at critical moments, effectively integrating asynchronous and cross-device response information. Furthermore, we construct multi-device datasets of thirty users in real-world scenarios to validate the effectiveness of VIP. Extensive experiments demonstrate that VIP performs well on different devices, with an average authentication accuracy improvement of 5.65% compared to existing methods. In addition, the false acceptance rates of VIP under simulated attacks and replay attacks are 1.55% and 2.71%, respectively, which are on average 0.61 and 0.28 percentage points lower than existing methods.
Keywords
dual-terminal collaboration, user authentication, vibration signal
Document Type
Journal Article
Date of Publication
1-1-2026
ISSN
15455971
Publication Title
IEEE Transactions on Dependable and Secure Computing
Publisher
IEEE
School
School of Business and Law
Copyright
subscription content
Content Type
Metadata only
Recommended Citation
Cui, Z., Yang, Z., He, Z., Kong, C., He, J., Li, J., Wang, P., & Liu, Z. (2026). Can vibration patterns identify users? Authentication for smartphone-watch collaboration. IEEE Transactions on Dependable and Secure Computing. Advance online publication. https://doi.org/10.1109/TDSC.2026.3700835