Abstract
This paper proposes a new method for improving an existing abnormality detection system for person who soaks in a bathtub. As the number of aged people increases year by year in Japan, bathing accident of the aged is growing at a rapid rate, especially in-bathtub drowning accident. Therefore, prompt detection of bather's abnormality such as dizziness and fainting is important to prevent in-bathtub drowning. In order to detect bather's abnormality promptly, an abnormality detection system using seven ultrasonic sensors has been proposed. The system uses the following two methods: posture detection and behavior detection, to detect bather's different state from normal before an accident occurs, and improves a delay of detection considered to be a serious problem heretofore. There was however plenty of room for improvement. In order to improve detection rate of the system, we propose a new detection method in this paper. The method uses two ultrasonic sensors to beam bather's head and neck, and detects the head height and swing speed of the head. Experimental results are superior to the accuracy of the existing system, which enables us to detect bather's abnormality more accurately.
