Abstract: In order to satisfy the requirements of the elderly care, and lessen the physical and psychological hurt caused by falling, the authors presented a fall detection system using a tri-axis accelerometer. The study idea was based on the attitude estimation using tri-axis accelerometer for the judgment of falling detection. In addition, considering the impact from the noise and the high accuracy of the falling detection system, the Kalman filtering algorithm was used to improve the system's reliability. The experimental results show that the system can alarm 100% when the human body is falling fore-and-aft, laterally and rapidly rising after falling, and can achieve the detection level of normal human falling.

Key words: fall detection, tri-axis accelerometer, attitude angle, Kalman filtering, telecare

摘要： 为了满足老年人的护理需求，减少老年人因跌倒造成的身心伤害，提出了一种基于三轴加速度传感器的人体跌倒检测系统。该系统主要基于姿态测量特性，利用姿态角作为跌倒判断标准；并且考虑到噪声影响和跌倒检测系统对检测正确率的高要求，利用Kalman滤波算法来提高算法精确度。实验结果表明该系统在人体前后、侧向跌倒和跌倒后迅速站起的情况下可以100%报警，达到人体正常跌倒情况的检测标准。

关键词: 跌倒检测, 三轴加速度传感器, 姿态角, 卡尔曼滤波, 远程监护