基于iOS的老年人跌倒检测报警系统研究

目的:设计一种基于手机的老年人跌倒检测报警系统,为跌倒后的老人提供及时的帮助,争取更多的急救时间.方法:利用iPhone内置的三轴加速度传感器和陀螺仪提取人体加速度和角速度数据,计算与人体运动相关的特征值信号向量幅值(signal vector magnitude,SVM)、信号幅值面积(signal magnitude area,SMA)、倾斜角(tilt angle,TA),采用多阈值判断算法判别人体是否发生跌倒.发生跌倒时,手机向监护人发出跌倒报警.结果:在系统准确性检测实验中,模拟老年人行走、慢跑、坐下、躺下、弯腰及前向跌倒、侧向跌倒、后向跌倒.慢跑识别准确率为96％,出现错报;前向跌倒检测准确率为98％,出现漏报;其余准确率皆为100％,系统正确率为99.25％.结论:该系统可直接利用手机内置传感器有效检测跌倒,并且对手机的放置方位无要求,是一种易于接受且更为可行的跌倒检测系统.

IOS-based alarm and detect system of falls in elder

Objective To present a phone-based fall detection system that can provide timely help for the elderly when a fall happens and get more time for the emergency if injuries occur due to falls. Methods According to the study, body acceleration and angular velocity data, obtained from the tri-accelerometer and gyroscope in iPhone, were used to calcu- late the eigenvalues of SVM, SMA and TA which were associated with the body movement. Multi-threshold algorithm was utilized to determine whether a fall would occur. A fall alert was sent to the guardian in the preset means in case of fall. Results The daily activities of the elderly such as walking, jogging, sitting, lying down, stooping and some kinds of fall like forward fall, sideways fall, back fall were chosen to test the accuracy of the fall detection algorithm. Among them, the accuracy of jogging recognition was 96%, the accuracy of forward fall detection was 98%, and the remained actions had 100 percent accuracy rate, and the system accuracy was 99.25%. Concluslons Effective fall detection with no additional inserted devices and no requirements for the placement of phones highlights the function and usage of this system.