ระบบตรวจจับและวิเคราะห์การนอนในท่าผิดปกติสำหรับผู้สูงอายุโดยใช้เทคโนโลยีสิ่งอำนวยความสะดวก

This research work, Sleep Irregularities Monitoring and Data Analytics for Elderly Persons using Assistive Technology, aims to develop an assistant system for monitoring elderly who have slept or spent on the bed for a long time. Taking care this kind of persons is one of the most important issue because it can lead to the elderly person’s health. For example, some elderly tends to sleep harder because of insomnia. This can cause hypertension, arrhythmia or Heart attack. Moreover, sleeping or spending most of time on bed with less changing the posture is not good for the elderly, who are bedridden. At the end of this research, the results can be used for in/direct in term of elderly care.

To achieve the ultimate goal, two kinds of bed sensor were designed and implemented in this research that were a large force sensor and a combination of force sensor and piezo sensor. These two kinds of sensors have different purposes to use. For a large force sensor, it is suitable for detecting the position of force on bed, and recognizing how long the force is applied on that position. Thus, the results of this sensor can be applied in the warning system for bed ridden. For a combination of force sensor and piezo sensor, they were designed for classifying the sleep posture. It can be used for analyzing the sleep behavior in elderly, and can be used in the warning system before the elderly is out of bed.

In this research, Artificial Intelligent (AI) was conducted for analyzing five postures on the bed which are “Lying down”, “Lying Left”, “Lying Right”, “Sitting”, and “Out of bed”. A combination between Neural Network and Bayesian Network was implemented as a main classification model. The experimental results yielded the maximum accuracy up to 94.65%. The combined model could improve the result of posture estimation and distinguish similar posture.
Furthermore, this research work was deployed at Hospital for Tropical Diseases, Faculty of Tropical Medicine, Mahidol University, and Banphaeo General Hospital for analyzing the bed sensor’s signal. In installation, a group of Alzheimer patients was selected to be a test case of this research because Alzheimer disease was often found in elderly.
Moreover, the warning system for bed fall prevention was also deployed for improving the quality of elderly care.