摘要
提出一种基于信道状态信息(Channel State Information,CSI)的室内人数统计方法。该方法通过商用无线设备,获取无线CSI数据,并对数据进行相位矫正和滤波处理,提取出有效且稳定的CSI幅度和相位信息。当人数增多时,CSI幅度和相位的离散程度和波动范围增大。以此为据,对幅度和相位进行特征提取后分类,统计出室内的人员数量。实验结果表明,该方法在多种分类算法情况下均有良好的表现,其中基于单层决策树的AdaBoost算法效果最佳,人员静止和随机运动时的准确度约为99.2%和96.4%,能准确统计出室内的人员数量。
This paper proposes a method for indoor people counting based on channel state information(CSI).This method obtained wireless CSI data through a commercial wireless device,performed phase correction and filter processing on the data,and then extracted effective and stable CSI amplitude and phase information.When the number of people increased,the dispersion and fluctuation range of CSI amplitude and phase increased.Based on this,the amplitude and phase were classified after feature extraction,and the number of indoor people was counted.The experimental results show that the proposed method has good performance under various classification algorithms,among which AdaBoost algorithm based on single-layer decision tree has the best effect,with the accuracy of 99.2%and 96.4%when people are stationary and moving randomly,which can accurately count the number of indoor people.
作者
史文华
戴欢
周泽仑
蒋敬敬
Shi Wenhua;Dai Huan;Zhou Zelun;Jiang Jingjing(College of Electronic and Information Engineering,Suzhou University of Science and Technology,Suzhou 215000,Jiangsu,China)
出处
《计算机应用与软件》
北大核心
2022年第11期105-111,共7页
Computer Applications and Software
基金
国家自然科学基金项目(61702354,61876121)
苏州科技大学科研项目(XKZ2017004)
江苏省物联网移动互联技术工程重点实验室开放课题(JSWLW2017004)
研究生科研创新计划项目(SKSJ18_012,SJCX19_0963)。
关键词
人数统计
信道状态信息
相位矫正
异常值检测
主成分分析
People counting
Channel state information
Phase correction
Abnormal value detection
Principal components analysis
