Toward Robust Photoplethysmography-Enabled Acute Mountain Sickness Risk Monitoring With Wearable Electronics

Wearable electronics and the Internet of medical things have enabled the availability of continuous physiological measurements, which provide a practical basis for reliable health monitoring. In this work, we address the problem of acute mountain sickness (AMS) risk monitoring at high altitudes, based on photoplethysmography (PPG) and altitude measurements collected from wrist-worn devices. To this end, a dual-branch Gaussian process framework is proposed to model the environmental and physiological effects on AMS risk through separate but coordinated branches, in which several PPG features are extracted in time and frequency domains in order to reflect the physiological characteristics from different perspectives. An uncertainty-aware scheme is further developed to enable reliable adaptive branch integration under personalized update settings. Performance evaluation conducted on a dataset obtained at high altitudes illustrates the effectiveness and robustness of the proposed approach compared with other baseline approaches, which demonstrates its potential for AMS risk monitoring under environmental–physiological mismatch scenarios.