Skip to Main Content
The elder population are at high risk of having a falling accident. Once it happens, timely assistance can dramatically reduce the damage. Obviously, real-time and accurate fall detection is the key. The algorithm proposed in this work can distinguish falls from ADL (activities of daily living) and give the information of both fall direction and lying posture through analyzing data from a accelerometer placed in the head level. The experimental results demonstrate that the algorithm is of important practical value.