Go to OpenReview Public Article ORCID homepageLife-Vest Embedded Ballistocardiogram Measurement System for Underwater Heart Rate and Heart Rate Variability Monitoring
Abstract: Seamless monitoring of vital signs in stranded persons could enhance rescue operations by providing physiological status and prioritizing efforts accordingly. In this work, we explore the feasibility of a skin-contact-free system embedded within a life vest for heart rate (HR) and HR variability (HRV) monitoring based on ballistocardiogram (BCG). The system is validated using human subject experiments performed both on land and underwater, and at different temperatures, against an electrocardiogram (ECG)-based gold-standard reference. Three different state-of-the-art processing pipelines were adapted to extract heartbeat timings. Multiple standard metrics were used to assess their performance; the adapted method based on Bruser et al. outperformed two other explored methods. HR correlation coefficients achieved by the best method were 0.92 (baseline on land), 0.90 (cold pressor on land), 0.87 (cold pool), and 0.96 (hot tub) [p < 0.001]. Bland-Altman analysis showed tight limits of agreement (LoA = ± 1.96 SD) across all environments, ranging from 2.37 bpm (hot tub) to 4.95 bpm (cold pressor on land). The detection rate reached ≥67.82% (pool), with a modified efficiency of ≥79.35% (pool). Different HRV metrics in the time and frequency domains were calculated. Mean interbeat intervals (mean IBIs) and standard deviation of successive normal-to-normal (SDNN) intervals outperformed other metrics, in terms of mean absolute error (MAE) (≤ 12.77 ms and ≤16.28 ms, respectively) and LoA (. 36.48 ms). This work presents significant progress in underwater cardiac physiology monitoring without ECG, demonstrating the feasibility of a BCG-based monitoring system integrated into a life vest to aid in rescue operations.
External IDs:doi:10.1109/jsen.2025.3618770
Loading