SinkNet: Interactive Sink to Detect Living Habits for Healthcare and Quality of Life Using Private Networks

With the development of IoT applications, a wide variety of healthcare systems are available. A decrease in birthrate and an aging population have led to increased importance of health care for older adults. In such environments, a monitoring system is useful, in which a targeted person can be continuously viewed through a camera, and reports regarding the person's health can be automatically transmitted through email. However, the application does not take security into account enough because most systems are connected to the Internet. The system should be protected from a wide variety of risks to keep both personal data and sensors safe. Therefore, we developed SinkNet. This program connects several sensors and constructs a private network that can automatically perform a series of actions for client devices, such as collecting logs, detecting problems through machine learning, and checking to avoid and solve problems. In addition, SinkNet manipulates multiple sensors and recognizes living habits (e.g., hand and face washing, gargling) from the sound of water flow and talks interactively to users through synthetic sounds. This paper describes SinkNet, focusing on its implementation and initial evaluation.