摘要
为了降低住院患者跌倒产生的严重后果,本文设计一种腕部穿戴的实时跌倒检测系统,系统以微处理器NRF52840为核心,采用惯性传感器LSM6DSOX和气压计BMP390采集数据。本文提出一种基于合角速度判断设备静止时刻来校准气压计估计的高度变化的方法,获得了更准确的高度变化数据。在算法上,本文提出改进的事件检测方法和决策树模型结合的方法,提高了跌倒检测的准确率。设备最终在床上跌倒、向前跌倒等4种跌倒和走、鼓掌、侧平举等9种日常行为的分类中实现了97.94%的准确度,可帮助护士实现对跌倒住院患者及时救助。
In order to reduce the serious consequences of inpatient falls,a wrist-worn real-time fall detection system with a microprocessor NRF52840 as the core is designed and an inertial sensor LSM6DSOX and a barometer BMP390 is used to collect data.a method is proposed to calibrate the height change estimated by the barometer based on the ensemble angular velocity to determine the device’s moment of rest,and more accurate height change data are obtained.In terms of algorithm,a method of improved event detection method and decision tree model is proposed to improve the accuracy of fall detection.The device finally achieves 97.94%accuracy in the classification of 4 types of falls such as bed fall and forward fall and 9 types of daily behaviors such as walking,clapping and side lifting,which can help nurses achieve timely assistance to fallen inpatients.
作者
乔星舒
刘亚东
廖成奥
刘晓亮
赵欣
王京华
QIAO Xingshu;LIU Yadong;LIAO Chengao;LIU Xiaoliang;ZHAO Xin;WANG Jinghua(College of Mechanical and Electric Engineering,Changchun University of Science and Technology,Changchun 130022,China;Department of Hematology,the First Hospital of Jilin University,Changchun 130021,China;Department of Pediatric Respiratory,the First Hospital of Jilin University,Changchun 130021,China)
出处
《传感器与微系统》
北大核心
2025年第3期111-116,共6页
Transducer and Microsystem Technologies
基金
国家“111”计划资助项目(D17017)
吉林大学科技成果概念验证项目(2024GN0011)。
关键词
跌倒检测
事件检测
气压计
决策树
腕部穿戴设备
fall detection
event detection
barometer
decision tree
wrist wearable device
