MagKey: Empowering Wearables with Ballistocardiography-based Key Generation through Magnetic Field Vibration Sensing

Symmetric key generation based on biometrics has emerged as a promising solution for wearables pairing. Among various biometrics, heartbeats offer significant potential owing to their inherent randomness and spontaneity. Ballistocardiography (BCG), in particular, stands out for its accessibility and inclusivity, as it measures the body's recoil forces in response to cardiac blood ejection into the vasculature. However, traditional approaches to BCG suffer from challenges in sensing on wearables and limited key generation rates. To this end, this paper presents MagKey, a system that enables wearables with BCG-based key generation. MagKey overcomes the difficulties in effective BCG sensing by translating skin vibration caused by recoil forces into magnetic field vibration (MFV). Moreover, MagKey demonstrates that the peak-to-peak trend (PPT) of MFV signals can reliably extract keys, and thus improve the key generation rate. To mitigate the impact of noise and motion artifacts on key generation, MagKey employs analog filters and a peak screening method for signal processing. We implement MagKey on a one-layer flexible printed circuit (FPC) and a two-layer printed circuit board (PCB). Extensive experiments show the usability and effectiveness of MagKey. Furthermore, our security analyses illustrate the scheme's resilience against potential attacks.