摘要
跌倒已逐渐成为一种危害老年人身体健康的严重事故,如何在跌倒前对跌倒作出预测具有重要意义。设计一种基于足底压力和惯性传感器的跌倒检测系统,系统位于鞋体外侧,同时设计一种三层BP神经网络作为检测算法,系统运行时采集传感数据并通过WiFi传给上位机,上位机对数据进行显示,特征处理后使用训练好的算法进行跌倒检测。实验结果表明,该系统对跌倒和日常活动(ADL)的准确度达99.7%,算法的敏感度和特异性分别为100%和99.3%,同时,检测系统的PIT值约为400ms。该系统在保证高准确率的同时,还实现了很高的PIT值,给跌倒后续处理保留了较长前置时间。
Falls have gradually become a serious accident endangering the health of the elderly.How to predict falls before they happen is of great significance.A fall detection system based on plantar pressure and inertial sensor is designed in this paper.The system is located on the outside of the shoe,and a detection algorithm based on three-layer BP neural network is designed.When the system is running,the sensor data is collected and transmitted to the upper computer via WIFI.The upper computer displays the data,and the trained algorithm is used for fall detection after feature processing.The experimental results show that the accuracy of the system for falls and daily activities(ADL)can reach 99.7%,the sensitivity and specificity of the algorithm are 100%and 99.3%,respectively.Meanwhile,the PIT value is about 400ms.the system not only ensures high accuracy,but also realizes extraordinarily high PIT value,which reserves a long lead time for subsequent treatment of falls.
作者
刘石雨
王多琎
LIU Shi-yu;WANG Duo-jin(Institute of Rehabilitation Engineering and Technology,University of Shanghai for Science and Technology;Shanghai Engineering Research Center of Assistive Devices;Key Laboratory of Neural-functional Information and Rehabilitation Engineering of the Ministry of Civil Affairs,Shanghai 200093,China)
出处
《软件导刊》
2021年第2期165-169,共5页
Software Guide
基金
上海市科技创新行动计划项目(19DZ2203600)。
关键词
跌倒检测
足底压力
加速度
角速度
BP神经网络
fall detection
plantar pressure
acceleration
angular velocity
BP neural network