基于多天线判决的CSI高效人体行为识别方法

doi:10.3778/j.issn.1673-9418.2005021

摘要/Abstract

摘要：

人体运动和行为分析成为普适计算中一个新兴的研究领域，针对目前行为识别方法成本高、精度低等问题，提出一种基于多天线联合判决的信道状态信息（CSI）高效人体行为识别方法（MADR）。所提方法分为三步：数据处理、特征提取、动作行为分类。首先，针对原始信号易受环境、设备干扰问题，该方法注重数据处理过程，分别使用Hampel、低通滤波器去除异常值、高频噪声，并进一步使用主成分分析去除带内噪声，得到平滑稳定的数据；其次，利用基于滑动方差的方式将包含时频域细节信息的第一主成分的无效信号进行剔除，得到有效表征行为动作的特征向量；最后，为充分利用多根天线的CSI特征，构建多个基于DTW的FKNN分类器在近邻样本级别上对行为动作进行联合判决。实验结果表明，该方法在会议室和实验室场景下的准确率分别为95.33%、92.67%，且与使用KNN分类器相比，大大缩短了系统训练时间。

关键词: WiFi信道状态信息, 多天线联合判决, 行为识别, 快速K近邻（FKNN）

Abstract:

Human motion and behavior analysis has become a new research field in pervasive computing. Aiming at the problems of high cost and low accuracy of current behavior identification methods, an efficient method MADR (multi-antenna joint decision efficient behavior recognition system) of CSI (channel state information) based on multi-antenna joint judgment is proposed. The method is divided into three steps: data processing, feature extraction, action and behavior classification. Firstly, to solve the problem that the original signals are susceptible to interference from environment and equipment, this method focuses on the data processing process. Hampel and low-pass filter are used to remove outliers and high-frequency noise, and principal component analysis is further used to remove in-band noise, so as to obtain smooth and stable data. Secondly, the invalid signals of the first principal component containing time-frequency domain details are eliminated by using the method based on sliding variance, and the feature vectors that effectively represent the behaviors and actions are obtained. Finally, in order to make full use of the CSI features of multiple antennas, a number of DTW-based FKNN (fast K nearest neighbor) classifiers are constructed to jointly judge behaviors at the level of neighboring samples. Experimental results show that the accuracy of the method is 95.33% and 92.67% respectively in the conference room and the laboratory, and the system training time is greatly reduced compared with the KNN (K nearest neighbor) classifier.

Key words: WiFi channel state information, multi-antenna joint decision, behavior recognition, fast K nearest neighbor (FKNN)

陶志勇, 郭京, 刘影. 基于多天线判决的CSI高效人体行为识别方法[J]. 计算机科学与探索, 2021, 15(6): 1122-1132.

TAO Zhiyong, GUO Jing, LIU Ying. Efficient Human Behavior Recognition Method of CSI Based on Multi-antenna Judgment[J]. Journal of Frontiers of Computer Science and Technology, 2021, 15(6): 1122-1132.

参考文献

[1] LI H, CHEN X, DU H, et al. Wi-Motion: a robust human activity recognition using WiFi signals[J]. IEEE Access, 2019, 7: 153287-153299.
[2] PU Q, GUPTA S, GOLLAKOTA S, et al. Whole-home gesture recognition using wireless signals[C]//Proceedings of the 19th Annual International Conference on Mobile Computing and Networking, Miami, Sep 30-Oct 4, 2013. New York: ACM, 2013: 27-38.
[3] SIGG S, SCHOLZ M, SHI S Y, et al. RF-sensing of activities from non-cooperative subjects in device-free recognition systems using ambient and local signals[J]. IEEE Transactions on Mobile Computing, 2014, 13(4): 907-920.
[4] ABDELNASSER H, YOUSSEF M, HARRAS K A. WiGest: a ubiquitous WiFi-based gesture recognition system[C]//Proceedings of the 2015 IEEE Conference on Computer Communications, Hong Kong, China, Apr 26-May 1, 2015. Piscataway: IEEE, 2015: 1472-1480.
[5] SHANG J C, WU J. A robust sign language recognition system with multiple Wi-Fi devices[C]//Proceedings of the Workshop on Mobility in the Evolving Internet Architecture, Los Angeles, Aug 25, 2017. New York: ACM, 2017: 19-24.
[6] WANG W, LIU A X, SHAHZAD M, et al. Device-free human activity recognition using commercial WiFi devices[J]. IEEE Journal on Selected Areas in Communications, 2017, 35(5): 1118-1131.
[7] WANG G, ZOU Y, ZHOU Z, et al. We can hear you with Wi-Fi![J]. IEEE Transactions on Mobile Computing, 2016, 15(11): 2907-2920.
[8] WANG T, YANG D D, ZHANG S Q, et al. Wi-Alarm: low-cost passive intrusion detection using Wi-Fi[J]. Sensors, 2019, 19(10): 23-35.
[9] XIAO F, CHEN J, XIE X H, et al. SEARE: a system for exercise activity recognition and quality evaluation based on green sensing[J]. IEEE Transactions on Emerging Topics in Computing, 2020, 8(3): 752-761.
[10] ZOU H, ZHOU Y X, YANG J F, et al. FreeCount: device-free crowd counting with commodity WiFi[C]//Proceedings of the 2017 IEEE Global Communications Conference, Singapore, Dec 4-8, 2017. Piscataway: IEEE, 2017: 1-6.
[11] YANG Z, LIU Y H. Wi-Fi radar: from RSSI to CSI[J]. Communications of the CCF, 2014, 10(11): 55-60.
杨峥, 刘云浩. Wi-Fi雷达: 从RSSI到CSI[J]. 中国计算机学会通讯, 2014, 10(11): 55-60.
[12] ZOU H, ZHOU Y X, YANG J F, et al. FreeDetector: device-free occupancy detection with commodity WiFi[C]//Proceedings of the 2017 IEEE International Conference on Sensing, Communication and Networking, San Diego, Jun 12, 2017. Piscataway: IEEE, 2017: 14-18.
[13] QIAN K, WU C, YANG Z, et al. Enabling contactless detection of moving humans with dynamic speeds using CSI[J]. ACM Transactions on Embedded Computing Systems, 2018, 17(2): 1-18.
[14] ALI K, LIU A X, WANG W, et al. Recognizing keystrokes using WiFi devices[J]. IEEE Journal on Selected Areas in Communications, 2017, 35(5): 1175-1190.
[15] LIU L S, WEI Z C, ZHANG C H, et al. Lifetime estimation for human motion with WiFi channel state information[J]. Journal of Computer Applications, 2019, 39(7): 2056-2060.
刘立双, 魏忠诚, 张春华, 等. 基于WiFi信道状态信息的人员活动持续时间估计[J]. 计算机应用, 2019, 39(7): 2056-2060.
[16] LI H, YANG W, WANG J X, et al. WiFinger: talk to your smart devices with finger-grained gesture[C]//Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing, Heidelberg, Sep 12-16, 2016. New York: ACM, 2016: 250-261.
[17] LENG Y Q, ZHANG H W, ZHANG W, et al. Introduction to machine learning to actual combat: practical application of MATLAB[M]. Beijing: Tsinghua University Press, 2019.
冷雨泉, 张会文, 张伟, 等. 机器学习入门到实战: MATLAB实践应用[M]. 北京: 清华大学出版社, 2019.
[18] MELGAREJO P, ZHANG X Y, RAMANATHAN P, et al. Leveraging directional antenna capabilities for fine-grained gesture recognition[C]//Proceedings of the 2014 ACM Conference on Ubiquitous Computing, Seattle, Sep 13-17, 2014. New York: ACM, 2014: 541-551.