Intelligent Elderly Healthcare System Based On Internet Of Things

According to some social surveys,In China,the proportion of elderly people becomes larger and larger.Moreover,young people are apt to work and live in big cities,it is impossible for them to take care of their parents remotely.With the increasing of elderly population in China,the social elderly care problem is becoming more and more important.Currently,most nursing homes in China have some problems,such as poor facilities,unprofessional nurses and low informative levels.Furthermore,people don’t know what exact caring services are provided to their elders in nursing home.In order to solve these problems,a low-cost,adaptable and data-driven intelligent elderly care system is proposed in this paper.This system can overcome some disadvantages existing in the procedure of traditional care service supply,such as separate management,information isolation and the facilities can’t satisfy the requirement of real-time care delivery.The intelligent system consists of four parts: 1.The indoor positioning system based on RFID;2.The information management system;3.The Android App;4.The wellness condition forecasting system.The intelligent system has the ability to use the Internet of Things technologies to monitor and record the real-time location data of the elderly in a nursing home.The recorded daily location data can be transformed to daily activity data.We can use the daily activity information to establish models to detect anomalous states of the elderly.If the elderly behaves abnormally in the nursing home,the nurses could get alarms through App associated to the intelligent elderly care system,and precise care services could be delivered to the elderly.We conducted the testing experiment in Room 516,Feiyun Building,and Room 414,No,10 dormitory in Lanzhou University.There were 4 graduate students participating in our experiment.The intelligent system was used to predict whether their daily living pattern was regular by analyzing their location data.The experiment results indicated that it is feasible to predict whether the experiment participant’s behavior is regular by using this pattern.The forecasting accuracy of the ELM model was 90%,while the forecasting accuracy of the SVM model was 100%.