摘要
利用Wi-Fi信号中信道状态信息(Channel State Information,CSI)的变化特征可实现被动行为检测.通过预选择出CSI数据中性能优良的子载波和对环境区分度更好的MIMO天线对,实现对数据的深度预处理,以及在提取到幅值和相位差特征后,经过小波变换得到更细粒度特征的数据后处理算法,提高行为的识别率.实验结果表明,该算法在150组测试数据集上的分类准确率高达97.1%,比在同等条件下未经过预处理和后处理算法的分类准确率高约6.6%.
Passive behavior detection can be realized by using the change characteristics of channel state information(CSI)in Wi-Fi signals.By pre-selecting subcarriers with good performance and MIMO antenna pairs with better environmental discrimination,deep data preprocessing is realized,and after extracting amplitude and phase difference features,data post-processing algorithms such as fine-grained features are obtained through wavelet transform,so as to improve the recognition rate of behavior.Experimental results show that the classification accuracy of the proposed algorithm on 150 test datasets is as high as 97.1%,which is about 6.6%higher than the classification accuracy of the algorithm without the above pre-processing and post-processing under the same conditions.
作者
付裕
FU Yu(School of Electronic and Electrical Engineering,Anhui Sanlian University,Hefei 230601,China)
出处
《高师理科学刊》
2023年第5期34-38,58,共6页
Journal of Science of Teachers＇College and University
基金
安徽三联学院校级科研基金重点项目(KJZD2023003)
安徽省教育厅重点科研项目(KJ2021A1174)。
关键词
信道状态信息
小波变换
子载波选择
天线对选择
行为识别
channel state information
wavelet transform
subcarrier selection
antenna pair selection
behavior recognition
