Real-Time Sleep Apnea Diagnosis Method Using Wearable Device without External Sensors

Currently the diagnosis of sleep apnea is performed mainly in hospital by polysomnography. However, obstructive sleep apnea depend on various factors such as daily life pattern, sleep environment, and posture. Therefore, there is a need for a real-time wearable system that detects sleep apnea which is easy to use. In this paper, we suggest the sleep care system that can predict sleep apnea conveniently whenever wherever. We measured the respiration, SpO2, heartrate, and 3-ACC signals of sleep apnea patients using wearable device. We measured the respiration and SpO2 of patients to judge the levels of sleep apnea. Based on the measurement, we analyzed the heartrate and 3-ACC signals with various machine learning algorithms to determine if sleep apnea correlates with the measurement. As a result of this study, in real-time (640 mu s), we can diagnosis sleep apnea with 95% accuracy by only analyzing heartrate and 3-ACC signals in a typical smart watch without external sensors.