[1]饶　壮,丁大钊,王依菁.基于CSI主成分分割的人体动作识别方法[J].郑州大学学报(工学版),2025,46(06):49-57.[doi:10.13705/j.issn.1671-6833.2025.03.021]
　RAO Zhuang,DING Dazhao,WANG Yijing.Human activity recognition method ba＜x＞sed on CSI principal component segmentation[J].Journal of Zhengzhou University (Engineering Science),2025,46(06):49-57.[doi:10.13705/j.issn.1671-6833.2025.03.021]
点击复制

基于CSI主成分分割的人体动作识别方法()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
46
期数:
2025年06期
页码:
49-57
栏目:
出版日期:
2025-10-25

文章信息/Info

Title:
Human activity recognition method ba＜x＞sed on CSI principal component segmentation
文章编号:
1671-6833(2025)06-0049-09
作者:
饶　壮1;  丁大钊2;  王依菁2
1.郑州大学 网络空间安全学院,河南 郑州 450002;2.嵩山实验室,河南 郑州 450000
Author(s):
RAO Zhuang1;  DING Dazhao2;  WANG Yijing2
1.School of Cyber Science and Engineering, Zhengzhou University, Zhengzhou 450002, China; 2.Songshan Laboratory, Zhengzhou 450000, China
关键词:
信道状态信息;  活动分割;  卷积神经网络;  双向门控循环单元;  多头注意力机制
Keywords:
channel state information;  activity segmentation;  convolutional neural network;  bidirectional gated recurrent unit;  multi-head attention mechanism
分类号:
TN92TP183TP391.4
DOI:
10.13705/j.issn.1671-6833.2025.03.021
文献标志码:
A
摘要:
传统的基于信道状态信息(CSI)进行人体动作识别的方法具有输入数据冗余、提取特征单一等问题。基于此,提出了一种基于CSI主成分的双层滑动窗口机制分割的人体动作识别方法。首先,对振幅进行去异常值、降噪,对相位进行线性校准、降噪;其次,利用基于主成分分析的双层滑动窗口机制对预处理后的CSI数据进行活动分割,去除与运动无关的信息,提升模型的训练效率;再次,通过卷积神经网络与双向循环门控单元对CSI数据的空间和时间两个维度进行分析,并融合多头注意力机制聚焦关键信息,实现对人体动作的高精度识别;最后,在WiAR和BAHAR两个公开数据集上进行实验验证,结果表明:所提方法可以有效地对多种环境下的多种人体活动进行识别,并减少50%的数据量,WiAR数据集准确率达96.53%,极大改善了输入数据冗余和提取特征单一问题,所提方法优于其他现有的识别方法。
Abstract:
The traditional method of human activity recognition based on channel state information (CSI) suffers from issues such as input data redundancy and limited feature extraction. To address this, a human activity recognition approach based on CSI principal components and a dual-layer sliding window mechanism was proposed. Firstly, autlier removal and noise reduction were performed on the amplitude the use of a dual-layer sliding window mechanism based on principal component analysis enabled activity segmentation of preprocessed CSI data to eliminate irrelevant information and enhance model training efficiency. Subsequently, spatial and temporal analysis of the CSI data was conducted using convolutional neural network and bidirectional gated recurrent unit, with the integration of a multi-head attention mechanism to focus on key information for achieving high-precision recognition of human activities. Experimental validation was performed using the WiAR and BAHAR public datasets, demonstrating that the proposed method could effectively recognize various human activities in diverse environments, while reducing the data volume by 5%. The accuracy achieved on the WiAR dataset was 96.53%, indicating superior performance compared to existing methods.

参考文献/References:

[1]MOSLEH S, CODER J B, SCULLY C G, et al. Monitoring respiratory motion with Wi-Fi CSI: characterizing performance and the breathe smart algorithm[J]. IEEE Access, 2022, 10: 131932-131951. 

[2]TANG Z L, LIU Q Q, WU M J, et al. Wi-Fi CSI gesture recognition based on parallel LSTM-FCN deep space-time neural network[J]. China Communications, 2021, 18 (3): 205-215. 
[3]HAMEED H, USMAN M, TAHIR A, et al. Pushing the limits of remote RF sensing by reading lips under the face mask[J]. Nature Communications, 2022, 13: 5168. 
[4]SHALABY E, ELSHENNAWY N, SARHAN A. Utilizing deep learning models in CSI-based human activity recognition[J]. Neural Computing and Applications, 2022, 34(8): 5993-6010. 
[5]WANG J C, TIAN Z S, ZHOU M, et al. Leveraging hypothesis testing for CSI based passive human intrusion direction detection[J]. IEEE Transactions on Vehicular Technology, 2021, 70(8): 7749-7763. 
[6]BRENA R F, ESCUDERO E, VARGAS-ROSALES C, et al. Device-free crowd counting using multi-link Wi-Fi CSI descriptors in Doppler spectrum[J]. Electronics, 2021, 10(3): 315. 
[7]SOTO J C H, GALDINO I, CABALLERO E, et al. A survey on vital signs monitoring based on Wi-Fi CSI data [J]. Computer Communications, 2022, 195: 99-110. 
[8]KOTARU M, KATTI S. Position tracking for virtual reality using commodity Wi-Fi[C]∥2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2017: 2671-2681. 
[9]XU Q Y, HAN Y, WANG B B, et al. Indoor events monitoring using channel state information time series[J]. IEEE Internet of Things Journal, 2019, 6 (3): 4977-4990. 
[10] GU Y, REN F J, LI J. PAWS: passive human activity recognition based on Wi-Fi ambient signals[J]. IEEE Internet of Things Journal, 2016, 3(5): 796-805. 
[11] HALPERIN D, HU W J, SHETH A, etal. Tool release: gathering 802.11n traces with channel state information [J]. ACM SIGCOMM Computer Communication Review, 2011, 41(1): 53. 
[12] LI F M, AL-QANESS M, ZHANG Y, et al. A robust and device-free system for the recognition and classification of elderly activities[J]. Sensors, 2016, 16(12): 2043. 
[13] GUO L L, WANG L, LIN C, et al. Wiar: apublic dataset for Wi-Fi-based activity recognition[J]. IEEE Access, 2019, 7: 154935-154945. 
[14] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradientbased learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324. 
[15] GRAVES A, MOHAMED A R, HINTON G. Speech recognition with deep recurrent neural networks[C]∥2013 IEEE International Conference on Acoustics, Speech and Signal Processing. Piscataway: IEEE, 2013: 6645-6649. 
[16] MOSHIRI P F, SHAHBAZIAN R, NABATI M, et al. A CSI-based human activity recognition using deep learning [J]. Sensors, 2021, 21(21): 7225. 
[17] SHI Z G, ZHANG J A, XU R, et al. Deep learning networks for human activity recognition with CSI correlation feature extraction[C]∥ICC2019-2019 IEEE International Conference on Communications (ICC). Piscataway: IEEE, 2019: 1-6.
[18] SHANG S N, LUO Q Y, ZHAO J J, et al. LSTM-CNN network for human activity recognition using Wi-Fi CSI data[J]. Journal of Physics: Conference Series, 2021, 1883(1): 012139. 
[19] YAMAK P T, LIY J, GADOSEY P K, et al. A comparison between ARIMA, LSTM, and GRU for time series forecasting[C]∥Proceedings of the 2019 2nd International Conference on Algorithms, Computing and Artificial Intelligence.New York: ACM, 2020: 49-55. 
[20] MING X, CHENG W, ZHU R L, et al. Human activities recognition with amplitude-phase of channel state information using deep residual networks[C]∥2022 IEEE 17th Conference on Industrial Electronics and Applications (ICIEA).Piscataway: IEEE, 2022: 1-6.
[21] ALSAIFY B A, ALMAZARI M M, ALAZRAI R, et al. A dataset for Wi-Fi-based human activity recognition in line-of-sight and non-line-of-sight indoor environments [J]. Data in Brief, 2020, 33: 106534. 
[22] SHOWMIK I A, SANAM T F, IMTIAZ H. Human activity recognition from Wi-Fi CSI data using principal component-based wavelet CNN[J]. Digital Signal Processing, 2023, 138: 104056. 
[23] ZHOU Q Z, XING J C, YANG Q L, et al. Measuring intrinsic human activity information using Wi-Fi-based attention model[J]. Measurement, 2022, 195: 111084. 
[24] GUO L L, ZHANG H, WANG C, et al. Towards CSIbased diversity activity recognition via LSTM-CNN encoder-decoder neural network [J]. Neurocomputing, 2021, 444: 260-273.

相似文献/References:

[1]马栋榉,李 敏,张端金.基于波形边缘检测的Wi-Fi室内定位信号TOA估计[J].郑州大学学报(工学版),2022,43(03):15.[doi:10.13705/j.issn.1671-6833.2022.03.014]
　MA Dongju,LI Min,ZHANG Duanjin.TOA Estimation of Indoor Localization by Wi-Fi Signal Based on Waveform Edge Detection[J].Journal of Zhengzhou University (Engineering Science),2022,43(06):15.[doi:10.13705/j.issn.1671-6833.2022.03.014]

更新日期/Last Update: 2025-10-21