基于UWB二维信道状态信息的室内人员摔倒检测方法  

Indoor personnel fall detection method based on UWB 2D channel state information

作　　者：王萍 高娇娇 张振亚[1,2] 殷涛 王文凯 WANG Ping;GAO Jiaojiao;ZHANG Zhenya;YIN Tao;WANG Wenkai(Anhui Province Key Laboratory of Intelligent Building and Building Energy Saving,Anhui Jianzhu University,Hefei 230022,China;School of Electronics and Information Engineering,Anhui Jianzhu University,Hefei 230601,China)

机构地区：[1]安徽建筑大学智能建筑与建筑节能安徽省重点实验室,安徽合肥230022 [2]安徽建筑大学电子与信息工程学院,安徽合肥230601

出　　处：《传感器与微系统》2025年第2期155-159,共5页Transducer and Microsystem Technologies

基　　金：安徽省高校学科(专业)拔尖人才学术资助项目(gxyq2022030);安徽省高校学科(专业)拔尖人才学术资助项目(gxbjZD2021067);安徽省高校自然科学研究重点项目(KJ2020A0470);安徽省特支计划创新领军人才项目(皖组办[2022]21号);安徽建筑大学智能建筑与建筑节能安徽省重点实验室主任基金资助项目(IBES2022ZR01)。

摘　　要：针对现有摔倒检测方法存在的隐私保护度不够、准确率不高及可穿戴要求高等问题,提出一种基于超宽带(UWB)雷达和深度学习的摔倒检测方法。该方法使用一种商用UWB雷达信号采集板采集2D信道状态信息(CSI)数据。经小波去噪后,通过深度神经网络(DNN)模型进行人员摔倒检测。其中,2D CSI数据由快时间和慢时间2个维度构成,所构造的深度神经网络模型由卷积神经网络(CNN)、长短记忆(LSTM)网络及全连接神经网络(FCN)三部分组成。最后,在高校科研实验室环境下,开展了实测实验以验证方法性能。结果表明:使用UWB2D CSI数据对室内人员摔倒检测具有较好检测性能,检测准确率可达97%;通过实验探讨了2D CSI窗口步长和设备安装位置对方法性能的影响。Aiming at the problem of insufficient privacy otection,low accuracy and high wearabity requirements in the current fall detection methods,a fall detection method based on ultra-wideband(UWB)radar and deep learning is proposed.A commercial-on-shelf UWB radar signal acquisition board is adopted to collect 2D channel state information(CSI)data.After wavelet the noising,through,consist of fast time and slow time deep neural network(DNN)models detect whether there is a human fall happening,of which 2D CSI data.The deep neural network model is composed of three parts,convolutional neural networks(CNN),long short-term memory(LSTM)networks,and the full connected neural network(FCN).Finally,extensive experiments in a research laboratory environment in an University,are carried out to verify performance of the method.Results show that the USEING UWB 2D CSI data has good detection performance for indoor personnel fall detection,detection accuracy is up to 97%.Influence of the step length of 2D sliding window and the device installation position of on the performance of the proposed method is carefullydiscussed through experiments.

关 键 词：摔倒检测 超宽带雷达 二维信道状态信息 深度神经网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]