Time-bound Contextual Bio-ID Generation for Minimalist Wearables

TL;DR

This work tackles the lack of real-time authentication in minimalist wearables by introducing Proteus, a framework that generates time-bound contextual bio-IDs from IMU and PPG signals and embeds them into a universal latent space via contrastive learning. The approach enables instantaneous device-to-device and human-to-device verification without enrolment, using a server-trained global model and adaptive, sensor-aware deployment. Key contributions include a SimCLR-inspired embedding scheme for cross-device consistency, a binary bio-ID matching module, and comprehensive evaluation on the FatigueSet dataset showing $TPR$ around $83$–$89\%$ with manageable $FPR$/$FNR$ across activities and placements. The findings indicate potential for secure, dynamic collaboration among minimalist wearables, while highlighting limitations such as dataset size and privacy concerns, guiding future work toward larger studies and privacy-preserving embedding techniques. Overall, Proteus advances instant, context-aware authentication in pervasive wearables, enabling secure multi-device interactions without explicit user enrolment.”trimmed

Abstract

As wearable devices become increasingly miniaturized and powerful, a new opportunity arises for instant and dynamic device-to-device collaboration and human-to-device interaction. However, this progress presents a unique challenge: these minimalist wearables lack inherent mechanisms for real-time authentication, posing significant risks to data privacy and overall security. To address this, we introduce Proteus that realizes an innovative concept of time-bound contextual bio-IDs, which are generated from on-device sensor data and embedded into a common latent space. These bio-IDs act as a time-bound unique user identifier that can be used to identify the wearer in a certain context. Proteus enables dynamic and contextual device collaboration as well as robust human-to-device interaction. Our evaluations demonstrate the effectiveness of our method, particularly in the context of minimalist wearables.

Figures (3)

• Figure 1: Accelerometer from headband and wristband (left) and PPG from left and right earbuds (right).
• Figure 2: Example of PPG (left) and embeddings (right) on two earbuds. (A) Same user, time-aligned; (B) Different users; (C) Same user, non time-aligned.
• Figure 3: Bio-ID Generation (Left) and Matching (Right)