Autonomous Sleep Motion Recognition

Human sleep motion is closely related to health, especially for the weak and the sick, and real-time monitoring sleep motion can effectively prevent from diseases [1, 2]. Different types of rhythmic exercises may be accompanied by different health challenges. For example, frequent turnovers, which are common among the elderly, may lead to insomnia, depression, and physical injury. When the children or the elderly lie prone, the obstruction of the mouths and noses may lead to obstructive sleep apnea (i.e. asphyxia). However, the lateral position can help alleviating the damage of reflux gastritis and other gastric diseases toward the patients [3, 4]. Therefore, the autonomous sleep motion recognition (SMR) itself is not only necessary to detect sleep rhythmic motion but also to accurately identify different types of rhythmic motions, such as turning over, lateral lying, and prone lying, to help diagnose potential sleep disorders.