Easy-to-Deploy Living Activity Sensing System and Data Collection in General Homes
Tomokazu Matsui, Kosei Onishi, Shinya Misaki, Manato Fujimoto, Hirohiko Suwa, Keiichi Yasumoto · 2020
Emergence of smart appliances and high performance IoT devices is promoting studies on more functional and intelligent home services using these devices. Especially, in developed countries including Japan with aging population and declining birthrate, it is urgent to develop technologies to monitor living situations of residents including elderly persons and improve their quality of life (QoL) through home services based on the activity recognition technology. However, activity recognition systems in general require many types/number of sensors and hence they are difficult to deploy and operate. In this paper, we propose a system consisting of low-cost and easy-to-deploy sensors based on energy harvesting that can collect data of resident's activities of daily living (ADL) for months without maintenance. The system was deployed in 10 homes of senior citizens where we collected ADL data for two months each. We also estimated the ADLs from the collected data by using long short-term memory (LSTM), a deep learning model. As a result, ADLs could be estimated at high recall rate of 82.4% on average and hence we found that the proposed system has high applicability to actual services.