基于室内无线传感器网络射频信号的老年人跌倒检测研究  

Fall Detection Using Radio Signals of Home Wireless Sensor Networks

作　　者：霍宏伟[1,2,3] 张宏科[1] Youzhi XU 

机构地区：[1]北京交通大学下一代互联网互联设备国家工程实验室,北京100044 [2]延雪平大学工学院,数字通信实验室,瑞典延雪平SE55611 [3]中国科学技术交流中心,北京100045

出　　处：《电子学报》2011年第1期195-200,共6页Acta Electronica Sinica

基　　金：国家自然科学基金(No.60773182);北京交通大学优秀博士生科研创新基金(No.48022);EU Regional Strategic Fund(X2.304-1135-06)

摘　　要：利用无线信号的自然衰减,在不显著增加通信开销的基础上,提出了一种新的老年人跌倒行为的检测方法.给出阶段相关性这一概念并用以区分体域传感器网络节点与室内传感器网络节点信号在人运动与静止条件下的统计相关性.给出了最小通信决策集合的概念,通过对比最小通信决策集合的内容,提出了老年人位置估计方法和跌倒行为检测算法;利用仿真工具分析了该方法的通信开销.用Micaz节点实现了集中式检测方法并进行了实验,结果证明本方案具有较高的检测准确性.This paper proposes a fall detection approach to detect accidental falls for senior citizens using wireless sensor networks based health-care systems.Unlike traditional sensory-context based detection,this paper achieves this goal by using radio signals in wireless sensor networks.A group of the sensors on the senior citizens＇body（wearable sensors） and several sensors as anchor nodes（anchor sensors） in their daily territories are deployed.The fall event is detected by comparing the difference of signal strengths from wearable sensors caused by the natural characteristics of the wireless radio signals attenuation through human bodies.Both centralized and distributed lightweight algorithms are proposed to compute the status of the fall event from senior citizens＇normal daily behaviors.The obtained evaluation and experiment results using MicaZ motes show that this approach performs high accuracy,low overhead and less response delay in fall events detection and report.

关 键 词：无线传感器网络 跌倒检测 阶段互相关系数 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]