PI-PPG: Motion-Robust and Bias-Reduced Wearable PPG via Light Polarization

Shailabh Kumar, Jongboo Kim, InChun Lim, Sang Kyu Kim, Radwanul Hasan Siddique

Confirmation: I have read and agree with the IEEE BSN 2025 conference submission's policy on behalf of myself and my co-authors.

Keywords: photoplethysmography (PPG), polarization, skin-tone, bias, motion, HR, HRV, digital health, wearable sensors, wearable health

TL;DR: Improved cardiovascular feature estimation, reduced motion-artifacts, and reduced racial bias on wearable devices using novel polarization sensor-design.

Abstract: Photoplethysmography (PPG) is widely used in wearable health monitoring, but conventional PPG sensors remain vulnerable to motion artifacts, ambient light interference, and skin-tone–related signal degradation. We present PI-PPG, a polarization-integrated PPG sensor system that improves signal fidelity and fairness in wearable cardiovascular monitoring. By combining high-efficiency polarized illumination with reflection-enhanced, cross-polarized detection, PI-PPG significantly reduces superficial scattering and motion-induced noise. Experimental validation across a range of Fitzpatrick skin types under both resting and active conditions demonstrates substantial gains, including >200% reduction in ambient light interference, >100% improvements in perfusion index and AC SNR, and >20% enhancement in cardiovascular feature detection. Notably, the system improves Heart Rate (HR) and Heart Rate Variability (HRV) accuracy by 28% for darker skin tones and reduces motion artifacts by 22% compared to conventional PPG. These advancements underscore the potential of PI-PPG to enhance detection efficiency, reliability, and fairness in wearable health sensing, thereby promoting equitable digital health solutions.

Track: 2. Sensors and systems for digital health, wellness, and athletics

NominateReviewer:

Shailabh Kumar
[e-mail]

Submission Number: 80