Unobtrusive Sensing for Sleep Quality Monitoring and Assessment
Sleep related disorders are common diseases especially among the elderly. In this paper, we propose a simple unobtrusive sensing environment for monitoring the elderly's sleep-wake conditions to assess sleep quality. The environment excludes the all-possible intrusive or contact required sensors (such as audio, video and even any wearable sensors) and only contains fixed type motion sensors (PIRs), a semi-fixed type accelerometer sensor (attached to the bed) and a set of smart algorithms; thus, it is nonintrusive, comfortable and cost- effective and can be used for long-term sleep monitoring, detecting early symptoms of sleep related disorders, and responding to caregivers. We implement and test the environment in several different real daily living environments. Our study shows that the detecting sensitivity is improved up to 96.61%, specificity is 91.81% and the Area under ROC curve (AUC) performance is 94.21% using a multilayer perception network. This result indicates that fusing different data from multiple sensors can obtain more reliable performance in detecting sleep and wake states.