Gait recognition method based on multi-feature fusion using frequency-modulated continuous wave radar

doi:10.11772/j.issn.1001-9081.2025091139

Journal of Computer Applications ›› 2026, Vol. 46 ›› Issue (10): 3317-3323.DOI: 10.11772/j.issn.1001-9081.2025091139

• Multimedia computing and computer simulation • Previous Articles

Gait recognition method based on multi-feature fusion using frequency-modulated continuous wave radar

Jiawei JIANG, Jian GUO(), Mingwei YUAN, Zhichun LI, Lijuan SUN

School of Computer Science，Nanjing University of Posts and Telecommunications，Nanjing Jiangsu 210003，China

Received:2025-09-26 Revised:2025-11-21 Accepted:2025-11-24 Online:2025-12-02 Published:2026-10-10
Contact: Jian GUO
About author:JIANG Jiawei， born in 2001， M. S. candidate. Her research interests include machine learning， gait recognition， behavior recognition.
YUAN Mingwei， born in 2006. Her research interests include machine learning， gait recognition.
LI Zhichun， born in 1997， M. S. His research interests include machine learning， gait recognition， wireless sensing.
SUN Lijuan， born in 1963， Ph. D.， professor. Her research interests include internet of things， wireless sensing， biometric recognition.
Supported by:
National Natural Science Foundation of China(62272242)

基于多特征融合的调频连续波雷达步态识别方法

江嘉玮, 郭剑(), 袁铭蔚, 李智春, 孙力娟

南京邮电大学计算机学院，南京 210003

通讯作者: 郭剑
作者简介:江嘉玮（2001—），女，江苏南通人，硕士研究生，主要研究方向：机器学习、步态识别、行为识别
袁铭蔚（2006—），女，江苏南通人，主要研究方向：机器学习、步态识别
李智春（1997—），男，安徽滁州人，硕士，主要研究方向：机器学习、步态识别、无线感知
孙力娟（1963—），女，江苏南京人，教授，博士，CCF会员，主要研究方向：物联网、无线感知、生物特征识别。
基金资助:
国家自然科学基金资助项目(62272242)

Abstract

Abstract:

A gait recognition method based on multi-feature fusion using Frequency-Modulated Continuous Wave （FMCW） radar was developed to overcome the low recognition accuracy of the existing millimeter-wave point cloud gait recognition methods caused by relying on static spatial features excessively and lacking modeling of dynamic motion information. First， the three-dimensional coordinates of the point cloud were aligned spatially to obtain a unified spatial representation. On this basis， the velocity and signal-to-noise ratio features of target points were fused， and the spatial structure and dynamic information were modeled jointly. Second， cross-spatio-temporal gait features were further extracted by multi-scale convolution and pooling modules. Experimental results show that the proposed method achieves high recognition accuracy on multiple public datasets and a self-built dataset， with an average accuracy of 86.3% and a maximum accuracy of 92.2%. Across different scenarios， it has the recognition accuracy improved by 0.90% to 32.08% compared with comparative methods such as mmGaitNet and PointMLP. It can be seen that through multi-dimensional feature fusion and multi-scale modeling， the proposed method enhances the representation capability of gait dynamic features effectively， and improves the accuracy and robustness of millimeter-wave radar gait recognition.

Key words: gait recognition, multi-feature fusion, point cloud, Frequency-Modulated Continuous Wave (FMCW) radar, deep learning

摘要：

针对现有毫米波点云步态识别方法因过度依赖静态空间特征、缺乏动态运动信息建模，导致识别精度受限的问题，提出一种基于多特征融合的调频连续波（FMCW）雷达步态识别方法。首先，对点云的三维坐标进行空间对齐，以统一空间表示；在此基础上融合目标点的速度与信噪比特征，实现空间结构与动态信息的协同建模。然后，通过多尺度卷积与池化模块进一步提取跨时空步态特征。实验结果显示，所提方法在多个公开数据集与自建数据集上均取得了较高的识别精度，平均准确率为86.3%，最高达到92.2%；在不同场景下，相较于mmGaitNet和PointMLP等对比方法，它的准确率提升幅度为0.90%至32.08%。所提方法通过多维特征融合与多尺度建模有效增强步态动态特征的表达能力，从而提升毫米波雷达步态识别的精确性与鲁棒性。

关键词: 步态识别, 多特征融合, 点云, 调频连续波雷达, 深度学习

CLC Number:

TP391

Jiawei JIANG, Jian GUO, Mingwei YUAN, Zhichun LI, Lijuan SUN. Gait recognition method based on multi-feature fusion using frequency-modulated continuous wave radar[J]. Journal of Computer Applications, 2026, 46(10): 3317-3323.

江嘉玮, 郭剑, 袁铭蔚, 李智春, 孙力娟. 基于多特征融合的调频连续波雷达步态识别方法[J]. 《计算机应用》唯一官方网站, 2026, 46(10): 3317-3323.

Figures/Tables 15

References 25

[1]	Rani V， Kumar M. Human gait recognition： a systematic review［J］. Multimedia Tools and Applications， 2023， 82（24）： 37003-37037.
[2]	陈福仕，沈尧，周池春，等. 无监督学习步态识别综述［J］. 计算机科学与探索， 2024， 18（8）： 2014-2033.
	Chen Fushi， Shen Yao， Zhou Chichun， et al. Review of unsupervised learning gait recognition［J］. Journal of Frontiers of Computer Science and Technology， 2024， 18（8）： 2014-2033.
[3]	Chang Y， Wang L， Lin L， et al. Gait recognition of ME-SVM based on single accelerometer［J］. Journal of Mechanics in Medicine and Biology， 2023， 23（2）： No.2350029.
[4]	Kim M J， Han J H， Shin W C， et al. Gait pattern identification using gait features［J］. Electronics， 2024， 13（10）： No.1956.
[5]	Niculae A， Catruna A， Cosma A， et al. Gait recognition from highly compressed videos［C］// FG 2024. Piscataway： IEEE， 2024： 1-7.
[6]	Dang X， Tang Y， Hao Z， et al. PGGait： gait recognition based on millimeter-wave radar spatio-temporal sensing of multidimensional point clouds［J］. Sensors， 2024， 24（1）： No.142.
[7]	Li J， Li B， Wang L， et al. Passive multiuser gait identification through micro-Doppler calibration using mmWave radar［J］. IEEE Internet of Things Journal， 2024， 11（4）： 6868-6877.
[8]	Xue S， Du L， Shi Y， et al. Fine-grained spatial-temporal gait recognition network based on millimeter-wave radar point cloud［J］. IEEE Transactions on Geoscience and Remote Sensing， 2024， 62： No.5101316.
[9]	Huang Y， Wang Y， Shi K， et al. HDNet： hierarchical dynamic network for gait recognition using millimeter-wave radar［C］// ICASSP 2023. Piscataway： IEEE， 2023： 1-5.
[10]	魏永超，谢卫鑫，张娅岚，等. 基于多尺度特征融合的跨视角点云步态识别［J］. 电子测量技术， 2025， 48（10）： 109-116.
	Wei Yongchao， Xie Weixin， Zhang Yalan， et al. Cross-view point cloud gait recognition based on multi-scale feature fusion［J］. Electronic Measurement Technology， 2025， 48（10）： 109-116.
[11]	Wu J， Wang J. PCA-based SVM for automatic recognition of gait patterns［J］. Journal of Applied Biomechanics， 2008， 24（1）： 83-87.
[12]	Lu J， Zhang E. Gait recognition for human identification based on ICA and fuzzy SVM through multiple views fusion［J］. Pattern Recognition Letters， 2007， 28（16）： 2401-2411.
[13]	Ni Z， Huang B. Open-set human identification based on gait radar micro-Doppler signatures［J］. IEEE Sensors Journal， 2021， 21（6）： 8226-8233.
[14]	Ozturk M Z， Wu C， Wang B， et al. GaitCube： deep data cube learning for human recognition with millimeter-wave radio［J］. IEEE Internet of Things Journal， 2022， 9（1）： 546-557.
[15]	Janakaraj P， Jakkala K， Bhuyan A， et al. STAR： simultaneous tracking and recognition through millimeter waves and deep learning［C］// WMNC 2019. Piscataway： IEEE， 2019： 211-218.
[16]	Li T， Cao X， Liu H， et al. MTPGait： multi-person gait recognition with spatio-temporal information via millimeter wave radar［C］// ICPADS 2021. Piscataway： IEEE， 2021： 660-666.
[17]	Zhao P， Lu C X， Wang J， et al. mID： tracking and identifying people with millimeter wave radar［C］// DCOSS 2019. Piscataway： IEEE， 2019： 33-40.
[18]	Meng Z， Fu S， Yan J， et al. Gait recognition for co-existing multiple people using millimeter wave sensing［J］. Proceedings of the AAAI Conference on Artificial Intelligence， 2020， 34（1）： 849-856.
[19]	Wang C， Gong P， Zhang L. STPointGCN： spatial temporal graph convolutional network for multiple people recognition using millimeter-wave radar［C］// ICASSP 2022. Piscataway： IEEE， 2022： 3433-3437.
[20]	Piotrowsky L， Jaeschke T， Kueppers S， et al. Enabling high accuracy distance measurements with FMCW radar sensors［J］. IEEE Transactions on Microwave Theory and Techniques， 2019， 67（12）： 5360-5371.
[21]	Ester M， Kriegel H P， Sander J， et al. A density-based algorithm for discovering clusters in large spatial databases with noise［C］// KDD 1996. Menlo Park， CA： AAAI Press， 1996： 226-231.
[22]	Qi C R， Su H， Mo K， et al. PointNet： deep learning on point sets for 3D classification and segmentation［C］// CVPR 2017. Piscataway： IEEE， 2017： 77-85.
[23]	He K， Zhang X， Ren S， et al. Deep residual learning for image recognition［C］// CVPR 2016. Piscataway： IEEE， 2016： 770-778.
[24]	Ma X， Qin C， You H， et al. Rethinking network design and local geometry in point cloud： a simple residual MLP framework［PP/OL］. V2. arXiv （2022-11-29）［2025-08-17］..
[25]	Zhang R， Wang L， Wang Y， et al. Starting from non-parametric networks for 3D point cloud analysis［C］// CVPR 2023. Piscataway： IEEE， 2023： 5344-5353.

数据集	eps/m	分类准确率
数据集	eps/m	minPts=5	minPts=6	minPts=7	minPts=8
Room1	0.3	0.881	0.879	0.859	0.868
	0.4	0.899	0.903	0.863	0.877
	0.5	0.905	0.922	0.913	0.914
	0.6	0.896	0.850	0.884	0.847
Room2	0.3	0.726	0.762	0.770	0.730
	0.4	0.736	0.793	0.758	0.750
	0.5	0.744	0.807	0.799	0.790
	0.6	0.717	0.775	0.730	0.799
STPointGCN	0.3	0.581	0.703	0.752	0.603
	0.4	0.604	0.737	0.790	0.631
	0.5	0.624	0.821	0.782	0.691
	0.6	0.576	0.735	0.640	0.599
Selfdata	0.3	0.772	0.813	0.820	0.793
	0.4	0.846	0.850	0.831	0.839
	0.5	0.868	0.901	0.869	0.862
	0.6	0.843	0.863	0.853	0.805

数据集	eps/m	分类准确率
数据集	eps/m	minPts=5	minPts=6	minPts=7	minPts=8
Room1	0.3	0.881	0.879	0.859	0.868
	0.4	0.899	0.903	0.863	0.877
	0.5	0.905	0.922	0.913	0.914
	0.6	0.896	0.850	0.884	0.847
Room2	0.3	0.726	0.762	0.770	0.730
	0.4	0.736	0.793	0.758	0.750
	0.5	0.744	0.807	0.799	0.790
	0.6	0.717	0.775	0.730	0.799
STPointGCN	0.3	0.581	0.703	0.752	0.603
	0.4	0.604	0.737	0.790	0.631
	0.5	0.624	0.821	0.782	0.691
	0.6	0.576	0.735	0.640	0.599
Selfdata	0.3	0.772	0.813	0.820	0.793
	0.4	0.846	0.850	0.831	0.839
	0.5	0.868	0.901	0.869	0.862
	0.6	0.843	0.863	0.853	0.805

单帧点云中的点数	分类准确率
单帧点云中的点数	Room1	Room2	STPointGCN	Selfdata
32	0.342	0.662	0.542	0.620
48	0.869	0.741	0.735	0.824
64	0.922	0.807	0.821	0.901
128	0.843	0.639	0.624	0.646
144	0.828	0.632	0.588	0.580
160	0.832	0.624	0.569	0.554

单帧点云中的点数	分类准确率
单帧点云中的点数	Room1	Room2	STPointGCN	Selfdata
32	0.342	0.662	0.542	0.620
48	0.869	0.741	0.735	0.824
64	0.922	0.807	0.821	0.901
128	0.843	0.639	0.624	0.646
144	0.828	0.632	0.588	0.580
160	0.832	0.624	0.569	0.554

模型	模型组件				分类准确率
模型	空间坐标	速度	信噪比	特征对齐模块	Room1	Room2	STPointGCN	Selfdata
Model1	√				0.748	0.679	0.653	0.714
Model2	√	√			0.823	0.735	0.781	0.893
Model3	√		√		0.759	0.693	0.713	0.835
Model4	√	√	√		0.875	0.774	0.808	0.897
Model5	√	√	√	√	0.922	0.807	0.821	0.901

Gait recognition method based on multi-feature fusion using frequency-modulated continuous wave radar

基于多特征融合的调频连续波雷达步态识别方法

RichHTML

PDF

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 15

References 25

Related Articles 15

Recommended Articles

Metrics

方法	准确率
方法	Room1	Room2	STPointGCN	Selfdata
ResNet	0.708	0.611	0.702	0.772
DR	0.773	0.680	0.752	0.791
PointNet	0.748	0.645	0.745	0.714
MTPGait	0.747	0.661	0.773	0.875
mmGaitNet	0.876	0.709	0.780	0.860
PointMLP	0.901	0.757	0.795	0.893
Point-PN	0.913	0.741	0.806	0.877
本文方法	0.922	0.807	0.821	0.901

模型	参数量/10⁶	推理时间/ms
DR	0.690	13.72
ResNet	21.010	29.06
PointNet	1.600	94.51
MTPGait	5.370	22.81
mmGaitNet	1.330	31.03
PointMLP	13.230	1 546.67
Point-PN	0.906	1 703.26
本文方法	1.610	218.79

数据集	Acc-Mean	Acc-Std	F1-Mean	F1-Std
Room1	0.917	0.003	0.899	0.002
Room2	0.791	0.007	0.769	0.006
STPointGCN	0.813	0.006	0.792	0.004
Selfdata	0.895	0.004	0.879	0.003

[1]	Shang LIU, Zhaosen TANG, Hongyue LIU, Linfang DONG, Jin ZHOU. Multimodal sentiment analysis model for missing modalities under shared semantic conditions [J]. Journal of Computer Applications, 2026, 46(9): 2761-2768.
[2]	Yimin ZHOU, Wenhui FU, Jie LUO, Juan WANG. End‑to‑end robust video watermarking against screen‑recapturing attacks based on spatio-temporal feature enhancement network [J]. Journal of Computer Applications, 2026, 46(9): 2889-2897.
[3]	Wei CUI, Ping ZHANG, Zhuoling XIAO, Zhuohang CHEN. Incremental semantic mapping system based on visual SLAM [J]. Journal of Computer Applications, 2026, 46(9): 2800-2808.
[4]	Xinliang LIU, Yushi XU, Dubai LI, Yanzhao REN. Survey of knowledge graph-based question answering methods [J]. Journal of Computer Applications, 2026, 46(8): 2394-2410.
[5]	Ming LIU, Dongqi SHEN, Ziyang MENG. Point cloud registration network with dual-branch multi-level feature fusion [J]. Journal of Computer Applications, 2026, 46(8): 2584-2593.
[6]	Fengchun LIU, Xinying SHAO, Chunying ZHANG, Liya WANG, Jing REN. FCMdepth： monocular depth estimation framework with multi-scale feature optimization [J]. Journal of Computer Applications, 2026, 46(8): 2603-2611.
[7]	Wei LIU, Weigang LI, Zhiqiang TIAN. Representation learning method of hierarchical rotation-invariant geometric structure for point cloud classification and segmentation [J]. Journal of Computer Applications, 2026, 46(8): 2594-2602.
[8]	Xiaojin GUO, Xuyang SUI, Kenan ZHOU. Channel estimation of STAR-RIS-assisted hybrid-field communication system based on deep learning [J]. Journal of Computer Applications, 2026, 46(8): 2548-2554.
[9]	Jing LIU, Shaoze ZHAO, Xingang LIU, Haozhe NIU, Haipeng JI. Cross-condition microstructure data generation method for titanium alloys based on improved CGAN [J]. Journal of Computer Applications, 2026, 46(7): 2373-2382.
[10]	Miaogen LING, Rui JING, Wei FANG. Survey of research on applications of explainable deep learning in tropical cyclone forecasting [J]. Journal of Computer Applications, 2026, 46(7): 2318-2326.
[11]	Xiangyi WU, Hailiang YE, Feilong CAO. Point cloud completion method based on smooth-sharpen graph convolution [J]. Journal of Computer Applications, 2026, 46(7): 2267-2276.
[12]	Xiuli DU, Xing GAO, Xiaoyu ZHANG, Chengsheng PAN, Qijie ZOU. Video snapshot compressive imaging reconstruction method based on dense spatio-temporal deformable attention [J]. Journal of Computer Applications, 2026, 46(7): 2288-2296.
[13]	Guoyou ZHANG, Hongyu NIE, Lihu PAN, Rundong LEI. Point-MLPBLS： point cloud semantic segmentation network based on MLP cascaded broad learning system [J]. Journal of Computer Applications, 2026, 46(7): 2259-2266.
[14]	Zhenkai XIONG, Mengjun XU, Yinyin SUN, Xin WANG. Maritime ship detection algorithm under complex weather environments based on enhanced YOLOv8 [J]. Journal of Computer Applications, 2026, 46(6): 1998-2006.
[15]	Yi DU, Mingjin XU, Jiayi KONG, Liyao WANG, Chen ZHAO. Low-rank adaptive parameter-efficient fine-tuning algorithm based on YOLOv11 [J]. Journal of Computer Applications, 2026, 46(6): 1738-1745.