웨어러블 디바이스에서의 3축 가속도 센서를 이용한 고 연령층 수면-각성 감지 알고리즘 개발

초록

Sleep is an essential factor for humans, and it is evaluated through sleep staging in polysomnography. However, polysomnography has limitations such as the first night effect. Therefore, sleep staging algorithms based on wearable devices have been studied. However, these algorithms have limitations in performance for the elderly. In this study, Sleep-wake detection algorithm was designed using the accelerometer sensor of a wearable device, targeting the elderly with an average age of 72.6 years. As a result, when applying features that consider the sleep characteristics of the elderly to the deep learning network, better performance was achieved. Three of the four estimated parameters to assess sleep quality were significant. This study is the first to propose a sleep-wake detection network specifically designed for the elderly, and it is expected to contribute to convenient sleep monitoring and personalized sleep quality assessment.

키워드