现代情报2025，Vol.45Issue(12)：155-166,12.DOI:10.3969/j.issn.1008-0821.2025.12.013
可穿戴数据分析驱动的运动障碍跌倒风险识别研究
Data-Driven Fall Risk Recognition for Motor Impairments Using Wearable Device Analysis
摘要
Abstract
[Purpose/Significance]This study investigates fall-risk behaviors among populations with motor dysfunc-tion,aiming to construct a framework for a personalized,real-time fall early-warning system.The goal is to provide valu-able reference for relevant stakeholders in the domain of proactive health management.[Method/Process]Multimodal sen-sor data were input into a PDR-CNN two-stage decoupled recognition network for feature extraction.This methodology facilitates real-time processing and high-accuracy recognition of human activity data.[Result/Conclusion]Experimental evaluations conducted on the public benchmark dataset UniMiB-SHAR demonstrated that the proposed PDR-CNN model achieved recognition accuracies 4.83%and 3.57%higher than the Plain-CNN and DepConvLSTM baseline models,respectively.This research confirms the superior performance of the PDR-CNN model in identifying abnormal human activities,including falls.It is capable of providing more reliable fall risk warning support for individuals with motor dys-function and offers new perspectives for advancing the transition from traditional to proactive health management paradigms.关键词
可穿戴设备/运动障碍/跌倒风险识别/深度学习/主动健康管理Key words
wearable devices/dyskinesia/fall risk identification/deep learning/active health management引用本文复制引用
滕起,杨钰婷,印玥,胡广伟,李硕,黄奇..可穿戴数据分析驱动的运动障碍跌倒风险识别研究[J].现代情报,2025,45(12):155-166,12.基金项目
国家社会科学基金重大项目"大数据驱动的城乡社区服务体系精准化构建研究"(项目编号:20&ZD154) （项目编号:20&ZD154）
"十四五"国家重点研发计划项目"主动健康服务数智化技术区域综合应用示范"(项目编号:2023YFC3605800) （项目编号:2023YFC3605800）
江苏省研究生科研与实践创新计划项目"主动健康管理视角下老年人的活动风险感知模式研究"(项目批准号:KYCX24_0100) （项目批准号:KYCX24_0100）
江苏省研究生科研与实践创新计划项目"人类心肺疾病本体的构建及应用"(项目批准号:SJCX23_0010). （项目批准号:SJCX23_0010）
