摘要
基于减轻老年人跌倒危害的目的,以及时有效的进行跌倒检测和跌倒救助,设计并实现了一款基于Android智能手机的跌倒检测系统,实时采集人体日常行为数据,对数据进行研究分析,提出了一种支持向量机(SVM)和决策树相组合的跌倒检测算法,该算法既能简单快捷的进行手机日常跌倒监护,又能准确有效的进行跌倒检测判断。采用两次跌倒检测判断的方法,在日常行为通过SVM算法判断是跌倒行为时,再进行第二次决策树跌倒判断,优化了日常跌倒检测,使准确率达到99%。
In order topromptly fall detection and fall relief of old people, a fall detection system based on Android phones is designed and implemented. By collecting real-time sensor data and analyzing the data, a fall detection algorithm based on Supported Vector Machine and Decision Tree is proposed. The algorithm can be simple and quick for mobile phone daily monitoring, and can accurately and effectively detect the fall. There are two steps to fall detect, the daily behavior of the first SVM algorithm is used to detect the daily behavior, while fall, and then the second Decision Tree fall detect, in this way it effectively optimize the detection and improve the accuracy to 99%.
出处
《电子设计工程》
2016年第17期51-54,共4页
Electronic Design Engineering
关键词
跌倒检测
数据采集
SVM
决策树
fall detection
data collection
SVM
decision tree