Vibrational, Acoustic and Inertial Sensing on Smart Eyewear for Physiological Monitoring

Smart eyewear (sEW) represents a promising platform for physiological monitoring, yet sensor selection remains an open design question. This work characterizes three sensors co-located on the nose pads of a sEW: a vibrational microphone (V2S), an acoustic microphone, and an inertial measurement unit (IMU), evaluated across five tasks (apnea, nasal breathing, eating crunchy and soft food, and drinking) in both quiet and noisy conditions in ten healthy subjects. The V2S achieved the highest accuracy (86.7%) in quiet scenario, while the acoustic microphone collapsed under noise (16.4%), confirming the advantage of contact-based sensing. V2S and IMUfusion yielded the optimal configuration (weighted F1=86.4%). Additionally, V2S, IMU and acoustic microphone were evaluated for heart rate (HR) estimation during apnea: V2S achieved a median error within the 5 bpm frequency resolution limit, outperforming both the IMU, which showed higher variability, and the acoustic microphone, which performed poorly.