A wrist -worn fall detection system using accelerometers and gyroscopes

This paper presents a system that utilizes wrist-worn motion sensing devices to detect falls. The sensing device consists of a tri-axis accelerometer and a three-axis gyroscope. The user wears two devices, one on each wrist. The device transmits collected data to the computer through Zigbee. The system then analyses the collected data and determines if a fall event has occurred. The system can rule out the non-fall events such as clapping, lying down, jumping, and clapping during jumping, and will not misrecognize them as fall events. Compared to other devices worn on the head or the chest, the sensing devices are worn on the wrists, and hence will not drop off easily. This also saves the user from the trouble of taking them down and then putting them on again when the user changes clothes. According to the experimental results, the average sensitivity and specificity of the detection system reached 95% and 96.7%, respectively. It performed better than another wrist-worn fall detection mechanism.