计算机应用研究2018，Vol.35Issue(3)：839-843,5.DOI:10.3969/j.issn.1001-3695.2018.03.041
一种基于时序分析异常数据的跌倒行为监测方法
Fall detection method based on time series analysis of abnormal data
摘要
Abstract
In view of the problem of false alarm caused by the fall in the fall detection,this paper proposed a new kind of fall detection method which was based on time series analysis of abnormal data.The method analysed acceleration signal time seties,then detected abnormal data by calculating the similarity between adjacent windows.Finally it classified different behaviors abnormal data by the classifier algorithm.The accuracy of falling detection is 95％,which is higher than general fall detection,at the same time,the false alarm rate is decreased by 5.3％.This method is good fall detection algorithm.关键词
行为识别/时序分析/异常监测Key words
behavior recognition/time series analysis/outlier detection分类
信息技术与安全科学引用本文复制引用
王忠民,张新平,梁琛..一种基于时序分析异常数据的跌倒行为监测方法[J].计算机应用研究,2018,35(3):839-843,5.基金项目
国家自然科学基金资助项目(61373116) （61373116）
陕西省科技统筹创新工程计划项目(2016KTZDGY04-01) （2016KTZDGY04-01）
西安邮电大学研究生创新基金资助项目(114-602080102) （114-602080102）
