AHNNet:Human Activity Recognition Based on Hybrid Neural Network Combining Attention Mechanism

doi:10.7652/xjtuxb202105014

您当前的位置：

首页 >

文章列表页 >

AHNNet:Human Activity Recognition Based on Hybrid Neural Network Combining Attention Mechanism

更新时间：2025-07-09

- AHNNet:Human Activity Recognition Based on Hybrid Neural Network Combining Attention Mechanism
- Vol. 55, Issue 5, Pages: 123-132(2021)
- 作者机构：
  
  1. 郑州大学信息工程学院,郑州,450000
  2. 郑州大学软件学院,郑州,450000
- 作者简介：
- 基金信息：
- DOI：10.7652/xjtuxb202105014
  CLC： TP391
- Online First：10 May 2021，
  
  Published：2021
- 稿件说明：
移动端阅览
AHNNet:Human Activity Recognition Based on Hybrid Neural Network Combining Attention Mechanism[J]. 2021, 55(5): 123-132.
DOI：

AHNNet:Human Activity Recognition Based on Hybrid Neural Network Combining Attention Mechanism[J]. 2021, 55(5): 123-132. DOI： 10.7652/xjtuxb202105014.

摘要

针对Wi-Fi信号的行为感知研究中传统机器学习方法特征提取困难、深度学习方法特征提取方式单一

导致特征提取不充分、识别准确率不高等问题

提出融合注意力机制的人体行为识别混合神经网络模型AHNNet。在对信道状态信息影响因子分析的基础上

使用信道状态信息的振幅数据作为行为识别的基础数据; 采用时间滑窗将长时间人体活动序列分割为短时间序列

构建样本数据

克服全局人体行为数据非实时、长度不固定的缺点; 通过双向循环门控网络和时序卷积网络并行提取输入数据特征

充分挖掘数据潜在特征之间的关系; 在双向循环门控网络中融合注意力机制以强化数据特征

进一步提高模型性能; 将双向循环门控网络和时序卷积网络提取到的特征进行融合

增加特征的多样性; 将融合特征输入到Softmax分类器进行分类

得到人体活动数据对应的行为。与长短期记忆网络、双向循环神经网络进行了对比

实验结果表明:在标准数据采集室数据集上

AHNNet的行为识别正确率达到97.15%

比未使用注意力机制的模型分类正确率提高1.81%; 在公共数据集上

AHNNet的行为识别正确率比其他对比模型的提高至少0.65%

参数量下降47%; 在不同环境下

AHNNet在卧室环境中的正确率为95.7%

比标准数据采集室中的下降1.45%。AHNNet具有良好的识别效果和鲁棒性

并且在复杂的居家环境中应具有应用价值。

Abstract

Activity wireless sensing is an important technology to implement health monitoring. Although the research on activity perception based on Wi-Fi has made good progress

there are still some problems

such as the difficulty of feature extraction in the traditional machine learning and the single feature extraction method in deep learning

which leads to insufficient feature extraction and low recognition accuracy. Here AHNNet

a hybrid neural network of human activity recognition integrating attention mechanism is proposed. Analyzing the influence factors of channel state information

the amplitude data of channel state information are used as the basic data for activity recognition

and the time sliding window is used to divide the long time human activity sequence into short time series to construct the sample data

which solves the difficulties of non-real-time and non-fixed length global human activity data. AHNNet parallelizes the bidirectional gated recurrent network and the temporal convolutional network to extract the features of the input data

so as to fully reveal the relationship between potential features of data. To further improve the performance of model recognition

AHNNet is combined with attention mechanism to strengthen the main features of data in bidirectional gated recurrent network. Then

the features extracted from the bidirectional gated recurrent network and the temporal convolutional network are fused to increase the diversity of the features

and the fused features are input into the Softmax classifier for classification to obtain the activity corresponding to human activity data. Experimental results demonstrate that AHNNet has stronger classification ability than the other models

and the average accuracy rate achieves 97.15%

and AHNNet has fewer parameters while maintaining high accuracy. The competing model demonstrates that the accuracy rate is 95.7% in the bedroom environment

1.45% lower than that in the standard data acquisition room. AHNNet has good recognition effect and robustness

especially in complex home environment.

关键词

Keywords

references

GHASEMZADEH H, JAFARI R. Physical movement monitoring using body sensor networks: a phonological approach to construct spatial decision trees [J]. IEEE Transactions on Industrial Informatics, 2011, 7(1): 66-77.

WANG Yuxi, WU Kaishun, NI L M. WiFall: device-free fall detection by wireless networks [J]. IEEE Transactions on Mobile Computing, 2017, 16(2): 581-594.

WANG Tianben, WANG Zhu, ZHANG Daqing, et al. Recognizing parkinsonian gait pattern by exploiting fine-grained movement function features [J]. ACM Transactions on Intelligent Systems and Technology, 2016, 8(1): 1-22.

SHI Lei, ZHANG Yifan, CHENG Jian, et al. Two-stream adaptive graph convolutional networks for skeleton-based action recognition [C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2019: 12026-12035.

NGUYEN H T K, FAHAMA H, BELLEUDY C, et al. Low power architecture exploration for standalone fall detection system based on computer vision [C]∥Proceedings of the 2014 European Modelling Symposium. Piscataway, NJ, USA: IEEE, 2014: 169-173.

JIANG Du, ZHENG Zujia, LI Gongfa, et al. Gesture recognition based on binocular vision [J]. Cluster Computing, 2019, 22(S6): 13261-13271.

RAMAKIC' A, BUNDALO Z, BUNDALO D. A method for human gait recognition from video streams using silhouette, height and step length [J]. Journal of Circuits, Systems and Computers, 2020, 29(7): 2050101.

LARA Ó D, LABRADOR M A. A survey on human activity recognition using wearable sensors [J]. IEEE Communications Surveys and Tutorials, 2013, 15(3): 1192-1209.

GU Yu, WANG Yantong, LIU Zhi, et al. SleepGuardian: an RF-based healthcare system guarding your sleep from afar [J]. IEEE Network, 2020, 34(2): 164-171.

ZHAO Mingmin, LI Tianhong, ALSHEIKH M A, et al. Through-wall human pose estimation using radio signals [C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2018: 7356-7365.

Cisco Mobile. Cisco visual networking index: global mobile data traffic forecast update 2016-2021 [EB/OL]. [2020-09-01]. http:∥www.cisco.com/en/US/solutions/collateral/ns341/ns525/ns537/ns705/ns827/white_paper_c11-520862.pdf.

HALPERIRR D, HU W, SHETH A, et al. Tool release: gathering 802.11n traces with channel state information [J]. Computer Communication Review, 2011, 41(1): 53-53.

CAO Yangjie, WANG Fuchao, LU Xinxin, et al. Contactless body movement recognition during sleep via WiFi signals [J]. IEEE Internet of Things Journal, 2020, 7(3): 2028-2037.

LI Hong, YANG Wei, WANG Jianxin, 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. New York, USA: ACM, 2016: 250-261.

段鹏松, 周志一, 王超, 等. WiNet: 一种适用于无线感知场景的步态识别模型 [J]. 西安交通大学学报, 2020, 54(7): 187-195.

DUAN Pengsong, ZHOU Zhiyi, WANG Chao, et al. WiNet: a gait recognition model suitable for wireless sensing scene [J]. Journal of Xi'an Jiaotong University, 2020, 54(7): 187-195.

余星达, 陈文杰, 王鼎, 等. 非接触式身份识别的深度学习算法 [J]. 西安交通大学学报, 2019, 53(4): 122-127.

YU Xingda, CHEN Wenjie, WANG Ding, et al. A deep learning algorithm for contactless human identification [J]. Journal of Xi'an Jiaotong University, 2019, 53(4): 122-127.

WANG Fangxin, GONG Wei, LIU Jiangchuan. On spatial diversity in WiFi-based human activity recognition: a deep learning-based approach [J]. IEEE Internet of Things Journal, 2019, 6(2): 2035-2047.

YOUSEFI S, NARUI H, DAYAL S, et al. A survey of human activity recognition using wifi CSI [EB/OL]. [2020-09-01]. https:∥arxiv.org/pdf/1708. 07129.pdf.

CHEN Zhenghua, ZHANG Le, JIANG Chaoyang, et al. WiFi CSI based passive human activity recognition using attention based BLSTM [J]. IEEE Transactions on Mobile Computing, 2018, 18(11): 2714-2724.

LEA C, VIDAL R, REITER A, et al. Temporal convolutional networks: a unified approach to action segmentation [M]∥Lecture Notes in Computer Science. Cham, Germany: Springer, 2016: 47-54.

WU K S, XIAO J, YI Y W, et al. FILA: fine-grained indoor localization [C]∥Proceedings of the IEEE INFOCOM. Piscataway, NJ, USA: IEEE, 2012: 2210-2218.

HOCHREITER S, SCHMIDHUBER J. Long short-term memory [J]. Neural Computation, 1997, 9(8): 1735-1780.

CHUNG J, GULCEHRE C, CHO K, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling [EB/OL]. [2020-09-01]. https:∥arxiv.org/abs/1412.3555.

SCHUSTER M, PALIWAL K K. Bidirectional recurrent neural networks [J]. IEEE Transactions on Signal Processing, 1997, 45(11): 2673-2681.

DENIL M, BAZZANI L, LAROCHELLE H, et al. Learning where to attend with deep architectures for image tracking [J]. Neural Computation, 2012, 24(8): 2151-2184.

BAHDANAU D, CHO K, BENGIO Y. Neural machine translation by jointly learning to align and translate [EB/OL]. [2020-09-01]. https:∥arxiv.org/abs/1409.0473.

IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift [EB/OL]. [2020-09-01]. https:∥arxiv. org/abs/1502.03167.

SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al. Dropout: a simple way to prevent neural networks from overfitting [J]. Journal of Machine Learning Research, 2014, 15: 1929-1958.

LIN Min, CHEN Qiang. Network in network [EB/OL]. [2020-09-01]. https:∥arxiv.org/pdf/1312. 4400.pdf.

孙红帅,王霞,柳萱,等.频域注意力机制下的癫痫脑电信号分类.2021,55(2):129-135.[doi:10.7652/xjtuxb202102015]

谢雨杰,杜友田,张潇.面向新概念学习的图像描述生成模型.2020,54(12):37-44.[doi:10.7652/xjtuxb202012005]

张西宁,郭清林,刘书语.深度学习技术及其故障诊断应用分析与展望.2020,54(12):1-13.[doi:10.7652/xjtuxb202012 001]

张连超,乔瑞萍,党祺玮,等.具有全局特征的空间注意力机制.2020,54(11):129-138.[doi:10.7652/xjtuxb202011016]

张克旭,杜昌旺,赵浩淇,等.针对局灶性癫痫患者的脑电微状态分析.2020,54(9):157-163.[doi:10.7652/xjtuxb202009 018]

景海婷,张秦,陈曼,等.小样本域自适应的皮肤病临床影像识别方法.2020,54(9):142-148,156.[doi:10.7652/xjtuxb 202009016]

赵小强,梁浩鹏.使用改进残差神经网络的滚动轴承变工况故障诊断方法.2020,54(9):23-31.[doi:10.7652/xjtuxb 202009002]

张弘斌,袁奇,赵柄锡,等.采用多通道样本和深度卷积神经网络的轴承故障诊断方法.2020,54(8):58-66.[doi:10.7652/xjtuxb202008008]

段鹏松,周志一,王超,等.WiNet:一种适用于无线感知场景的步态识别模型.2020,54(7):187-195.[doi:10.7652/xjtuxb202007022]

孟月波,纪拓,刘光辉,等.编码-解码多尺度卷积神经网络人群计数方法.2020,54(5):149-157.[doi:10.7652/xjtuxb 202005020]

余艳杰,孙嘉琪,葛思擘,等.CycleGAN-SN:结合谱归一化和CycleGAN的图像风格化算法.2020,54(5):133-141.[doi:10.7652/xjtuxb202005018]

张帆,张景森,裴昌幸.采用双通道时间稠密网络的高光谱影像分类方法.2020,54(4):126-132.[doi:10.7652/xjtuxb 202004016]

刘天宇,周稻祥,李明,等.面向螺丝锁附序列的多分辨率融合卷积神经网络.2020,54(3):161-168,178.[doi:10.7652/xjtuxb202003020]

吴静然,丁恩杰,崔冉,等.采用多尺度注意力机制的旋转机械故障诊断方法.2020,54(2):51-58.[doi:10.7652/xjtuxb 202002007]

毛远宏,贺占庄,马钟,等.采用类内迁移学习的红外/可见光异源图像匹配.2020,54(1):49-55.[doi:10.7652/xjtuxb 202001007]

许美凤,国雷达,宋盼盼,等.多卷积神经网络模型融合的皮肤病识别方法.2019,53(11):125-130.[doi:10.7652/xjtuxb 201911018]

胡怀中,张建博,刘汉青,等.采用卷积神经网络的电站锅炉燃烧效率建模方法.2019,53(10):10-15.[doi:10.7652/xjtuxb201910002]

张康,陶建峰,覃程锦,等.随机丢弃和批标准化的深度卷积神经网络柴油机失火故障诊断.2019,53(8):159-166.[doi:10.7652/xjtuxb201908021]

贾晓芬,郭永存,柴华荣,等.深立井井壁图像的卷积神经网络去噪方法.2019,53(6):117-124.[doi:10.7652/xjtuxb2019 06016]

余星达,陈文杰,王鼎,等.非接触式身份识别的深度学习算法.2019,53(4):122-127.[doi:10.7652/xjtuxb201904018]

姜慧明,秦贵和,邹密,等.改进U型卷积网络的细胞核分割方法.2019,53(4):100-107,121.[doi:10.7652/xjtuxb2019 04015]

王晓芳,齐春.一种运用显著性检测的行为识别方法.2018,52(2):24-29,44.[doi:10.7652/xjtuxb201802004]

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Multimodal Neural Network-Based Temperature Prediction Method for Lithium-Ion Batteries Under Mechanical Abuse Scenarios

Application of Temporal Convolutional Network in Rotor Thermo-Mechanical Coupling Stress Prediction and Life Assessment

Research on Entity Relation Extraction Method Based on Relational Prompts with Single-Module Single-Step Approach

Cross-Component Prediction for H.266/Versatile Video Coding Based on Lightweight Convolutional Neural Network

Image Super Resolution Algorithm Based on Receptive Field Optimization and Progressive Feature Fusion

Related Author

HUANG Xinrong

QIN Zhengpeng

LI Qian

LI Jie

LI Pu

YUAN Qi

ZHANG Xiaolong

WEN Siguo

Related Institution

School of Energy and Electrical Engineering, Chang'an University

西安陕鼓动力股份有限公司，710075，西安Xi'an Shaangu Power Co., Ltd., Xi'an 710075, China

西安交通大学能源与动力工程学院，710049，西安School of Energy and Power Engineering, Xi'an Jiaotong University, Xi'an 710049, China

School of Computer Science and Technology, Harbin Institute of Technology

College of Electronic Engineering, National University of Defense Technology

⁰