跌倒事件已成為銀髮族中最常見的意外事故,近幾年來跌倒偵測系統的發展已因此引起廣泛關注。藉由跌倒偵測系統的協助,銀髮族可以即時獲得妥善醫療。本研究使用基於Android作業系統的智慧型手錶(WIMM One),藉由其內建之三軸重力加速計,並邀請六名受測者(3男3女),平均BMI為21,構建一個跌倒偵測系統。本研究於偵測平台整合出一套兩階段演算法,同時偵測線性和非線性活動。實驗結果顯示,靈敏度可提升到92.5%;特異性為95%,相較於先前演算法分別有12%與3%的提升。
Since fall event has become the most common accident occurred among elderly people, the development of fall detection system has received much attention in recent years. With the help of such a system, elderly persons can be delivered to a hospital to receive timely medical care. This study aims to implement a fall detection system by using an Android-based watch (the WIMM One) equipped with a tri-axial gravity accelerometer. Six young volunteers (3 males and 3 females) with an average BMI index of 21 were invited to conduct the experiments. By simultaneously considering detections for linear and non-linear movements, the proposed system achieves 92.5% in sensitivity and 95% in specificity, which indicated improvements of 12% and 3%, respectively, as compared to previous scheme.