一种结合多尺度小波与超球面表示的射频指纹开集识别方法

田欣玉; 李子睿; 郑庆河; 周福辉; 余礼苏; 黄崇文; 姜蔚蔚; 束锋; 赵毅哲

doi:10.11999/JEIT260214

一种结合多尺度小波与超球面表示的射频指纹开集识别方法

doi: 10.11999/JEIT260214 cstr: 32379.14.JEIT260214

1.
山东管理学院智能工程学院济南 250357
2.
南京航空航天大学人工智能学院南京 211106
3.
南昌大学信息工程学院南昌 330031
4.
浙江大学信息与电子工程学院杭州 310058
5.
北京邮电大学信息与通信工程学院北京 100876
6.
海南大学信息与通信工程学院海口 570228
7.
电子科技大学信息与通信工程学院成都 611731

基金项目: 国家自然科学基金(62401070)，山东省自然科学基金(ZR2019ZD01, ZR2023QF125)，山东省高等学校青年创新团队计划(2024KJH005)，山东省科技型中小企业创新能力提升工程(2024TSGC0055)

详细信息

作者简介:
田欣玉：女，讲师，研究方向为人工智能、模式识别、信号处理、物联网

李子睿：男，本科生，研究方向为无线通信、射频指纹识别、物联网、人工智能

郑庆河：男，教授，研究方向为无线通信、认知无线电、机器学习、调制识别

周福辉：男，教授，研究方向为电磁空间机器学习基础理论、认知智能与知识图谱、频谱智能共享和动态接入

余礼苏：男，副教授，研究方向为射频光载无线通信、非正交多址接入、无人机通信、人工智能

黄崇文：男，教授，研究方向为6G无线通信、智能协同感知、智能天线

姜蔚蔚：男，副教授，研究方向为卫星通信、无线通信、物联网、人工智能

束锋：男，教授，研究方向为智能无线通信、信息安全、大规模MIMO测向与定位

赵毅哲：男，副教授，研究方向为无线通信、通信控制一体化、流体天线

通讯作者:
郑庆河　zqh@sdmu.edu.cn

中图分类号: TN929.5
计量
- 文章访问数: 35
- HTML全文浏览量: 3
- PDF下载量: 4
- 被引次数: 0
出版历程
- 修回日期: 2026-03-27
- 录用日期: 2026-03-27
- 网络出版日期: 2026-04-22

A Radio Frequency Fingerprint Open Set Identification Method Combining Multi-Scale Wavelets and Hypersphere Representation

1.
School of Intelligent Engineering, Shandong Management University, Jinan 250357, China
2.
School of Artificial Intelligence, Nanjing University of Aeronautics and Astronautics, Nanjing 211106, China
3.
School of Information Engineering, Nanchang University, Nanchang 330031, China
4.
School of Information and Electronic Engineering, Zhejiang University, Hangzhou 310058, China
5.
School of Information and Communication Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, China
6.
School of Information and Communication Engineering, Hainan University, Haikou 570228, China
7.
School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China

Funds: The National Natural Science Foundation of China (62401070), The Shandong Provincial Natural Science Foundation (ZR2019ZD01, ZR2023QF125), The Shandong Provincial Youth Innovation Team Plan of Higher Education Institutions (2024KJH005), The Shandong Provincial Science and Technology Based Small and Medium sized Enterprises Innovation Capability Enhancement Project (2024TSGC0055)

摘要

摘要: 针对低信噪比环境下射频指纹特征易被噪声掩盖、多径效应引发非线性失真，以及现有方法在特征提取与未知设备检测能力上的不足，该文提出一种结合多尺度小波与超球面表示的射频指纹开集识别方法。首先，设计基于一维平稳小波变换的多尺度特征提取前端，实现对I/Q信号的全分辨率、多尺度分解，为后续网络提供高判别性输入。其次，构建多尺度残差注意力网络，融合深度残差学习、全局自注意力机制与双向长短时记忆网络，增强模型对微弱指纹特征的感知能力与长程时序依赖建模能力。最后，引入超球面度量学习，将特征空间约束至单位超球面，通过优化角度间隔构建类内紧凑、类间可分离的特征分布，并基于余弦相似度实现未知设备的有效检测。在基于IEEE 802.11协议的高保真仿真数据集上的实验结果表明，所提方法在–5 dB至20 dB信噪比范围内均能保持较高的开集识别准确率，平均分类准确率达65.34%，在–5 dB低信噪比下AUC达0.81，显著优于现有对比方法，验证了其在极端恶劣信道环境下的鲁棒性与有效性。
- 射频指纹识别 /
- 多尺度小波 /
- 超球面表示 /
- 多尺度残差注意力
Abstract: Objective To address the challenges of radio frequency fingerprint (RFF) identification in the open-set scenarios under low signal-to-noise ratio (SNR) conditions, where fingerprint features are easily masked by noise, multipath effects introduce nonlinear distortions, and existing methods struggle with feature extraction and unknown device detection, this paper proposes a deep learning framework that integrates multi-scale wavelet decomposition with hypersphere representation for robust open-set RFF identification. Methods The proposed method MS-RANet consists of three key components. First, a multi-scale feature extraction front-end based on one-dimensional stationary wavelet transform is designed to perform full-resolution and multi-scale decomposition of I/Q signals, preserving discriminative fingerprint information while suppressing noise. Second, a multi-scale residual attention network is constructed by incorporating deep residual learning, global self-attention mechanism, and bidirectional LSTM to enhance the model’s ability to perceive subtle fingerprint features and model long-range temporal dependencies. Finally, hyperspherical metric learning is introduced to constrain the feature space onto a unit hypersphere, optimizing angular margins to achieve compact intra-class and separable inter-class feature distributions. Unknown device detection is subsequently performed based on cosine similarity. Results and Discussions Extensive experiments conducted on high-fidelity IEEE 802.11-based simulation dataset demonstrate the effectiveness of MS-RANet. It achieves an average classification accuracy of 65.34% across SNR levels ranging from –5 dB to 20 dB, and an area under the curve (AUC) of 0.81 at –5 dB SNR, significantly outperforming existing methods such as DNN, GRU, CNN-LSTM, ResNet50, and DRSN-CA. Confusion matrices and receiver operating characteristic (ROC) curves further validate its robustness under extreme channel conditions. Feature visualization using t-SNE reveals that MS-RANet forms well-separated and compact clusters for known devices while effectively pushing unknown samples away from known class regions. Ablation studies confirm the individual contributions of multi-scale wavelet front-end, global attention, BiLSTM, and hyperspherical metric learning modules. Conclusions This paper presents a robust open-set RFF identification method that effectively combines multi-scale wavelet decomposition with hyperspherical representation learning. The proposed framework exhibits strong noise resilience, superior feature discrimination, and reliable unknown device detection under low-SNR and multipath fading conditions. The future work will focus on reducing computational complexity, improving inference speed, exploring generalization capability across different scenarios and protocols, and integrating the method with other physical-layer security mechanisms for collaborative authentication frameworks.
- Radio frequency fingerprint identification /
- Multi-scale wavelet /
- Hypersphere representation /
- Multi scale residual attention

HTML全文

图 1 通信系统信号模型

下载: 全尺寸图片幻灯片

图 2 结合多尺度小波与超球面表示的RFF开集识别架构

下载: 全尺寸图片幻灯片

图 3 多尺度特征提取前端结构

下载: 全尺寸图片幻灯片

图 4 多尺度残差注意力网络结构

下载: 全尺寸图片幻灯片

图 5 MS-RANet模型训练过程的收敛曲线

下载: 全尺寸图片幻灯片

图 6 典型信噪比下所提方法MS-RANet的混淆矩阵

下载: 全尺寸图片幻灯片

图 7 典型信噪比下各个方法的ROC曲线

下载: 全尺寸图片幻灯片

图 8 基于t-SNE的特征可视化对比(SNR = 20 dB)

下载: 全尺寸图片幻灯片

表 1 信号模型参数

参数类型	参数设置	参数类型	参数设置
通信协议	IEEE 802.11	帧类型	L-LTF
载波频率(GHz)	5	信道带宽(MHz)	20
采样长度(Points)	160	信噪比(dB)	–5:5:20
指纹参数$ \alpha $	[0.7, 1.3]	指纹参数$ \beta $	$ \approx \alpha -1 $
信道类型	多径瑞利衰落	直流偏移(dB)	[-50, -32]
设备数(已知/未知)	67/20	频率偏移(ppm)	[-4, 4]
样本划分(Train: Val: Test)	8:1:1	相位噪声(rad)	[0.01, 0.3]
样本总量(Frames)	87000	噪声类型	加性高斯白噪声

下载: 导出CSV

表 2 训练过程超参数

参数名称	数值	参数名称	数值
初始学习率	0.0008	最大迭代轮数	30
学习率衰减因子	0.6	丢弃率$ \rho $	0.3
学习率衰减周期	8	超球面半径$ s $	16
惩罚因子$ \lambda $	0.0001	特征空间维度$ d $	256
批量大小$ \Omega $	96	梯度裁剪阈值	1

下载: 导出CSV

表 3 RFF识别性能对比

对比模型	平均准确率(%)	AUC(SNR = –5 dB)	参数量(M)	计算量(GFLOPs)	推理时间(ms)
DNN^[20]	35.56	0.51	4.85	0.07	1.3
GRU^[21]	54.27	0.58	0.86	6.54	7.8
CNN-LSTM^[22]	54.35	0.67	1.15	8.12	6.4
ResNet50^[23]	56.44	0.62	23.5	4.13	9.6
DRSN-CA^[24]	58.92	0.73	1.65	2.45	7.2
本文MS-RANet	65.34	0.81	0.34	2.58	10.5

下载: 导出CSV

表 4 MS-RANet模块消融实验结果

序号	MS-FE 模块	BiLSTM 模块	全局注意力模块	HML 模块	平均准确率(%)	AUC (SNR = –5 dB)
1	×	×	×	×	35.61	0.53
2	√	×	×	×	57.85	0.68
3	√	√	×	×	59.23	0.70
4	√	√	√	×	60.15	0.72
5	√	√	√	√	65.64	0.79

下载: 导出CSV

参考文献(25)

[1]	HUAN Xintao, HAO Yi, MIAO Kaitao, et al. Carrier frequency offset in Internet of Things radio frequency fingerprint identification: An experimental review[J]. IEEE Internet of Things Journal, 2024, 11(5): 7359–7373. doi: 10.1109/JIOT.2023.3328025.
[2]	CHEN Yuchi. Fully incrementing visual cryptography from a succinct non-monotonic structure[J]. IEEE Transactions on Information Forensics and Security, 2017, 12(5): 1082–1091. doi: 10.1109/TIFS.2016.2641378.
[3]	ZHANG Zhentian, WONG K K, DANG Jian, et al. On fundamental limits for fluid antenna-assisted integrated sensing and communications for unsourced random access[J]. IEEE Journal on Selected Areas in Communications, 2026, 44: 136–149. doi: 10.1109/JSAC.2025.3608113.
[4]	LUO Hongyi, LI Guyue, BRIGHENTE A, et al. Channel-robust RF fingerprint identification for multi-antenna 5G user equipments[J]. IEEE Transactions on Information Forensics and Security, 2025, 20: 10761–10776. doi: 10.1109/TIFS.2025.3611154.
[5]	WANG Xuyu, WANG Xiangyu, and MAO Shiwen. RF sensing in the internet of things: A general deep learning framework[J]. IEEE Communications Magazine, 2018, 56(9): 62–67. doi: 10.1109/MCOM.2018.1701277.
[6]	ALSINDI N, CHALOUPKA Z, ALKHANBASHI N, et al. An empirical evaluation of a probabilistic RF signature for WLAN location fingerprinting[J]. IEEE Transactions on Wireless Communications, 2014, 13(6): 3257–3268. doi: 10.1109/TWC.2014.041714.131113.
[7]	LEE E, CHOI D H, NAM T, et al. An analysis of electromagnetic signatures from triangularly modulated spread spectrum clocking signals[J]. IEEE Transactions on Electromagnetic Compatibility, 2024, 66(3): 749–760. doi: 10.1109/TEMC.2024.3377245.
[8]	张在琛, 江浩. 智能超表面使能无人机高能效通信信道建模与传输机理分析[J]. 电子学报, 2023, 51(10): 2623–2634. doi: 10.12263/DZXB.20221352. ZHANG Zaichen and JIANG Hao. Channel modeling and characteristics analysis for high energy-efficient RIS-assisted UAV communications[J]. Acta Electronica Sinica, 2023, 51(10): 2623–2634. doi: 10.12263/DZXB.20221352.
[9]	LIN Yun, TU Ya, DOU Zheng, et al. Contour stella image and deep learning for signal recognition in the physical layer[J]. IEEE Transactions on Cognitive Communications and Networking, 2021, 7(1): 34–46. doi: 10.1109/TCCN.2020.3024610.
[10]	YIN Pengcheng, PENG Linning, ZHANG Junqing, et al. LTE device identification based on RF fingerprint with multi-channel convolutional neural network[C]. IEEE Global Communications Conference (GLOBECOM), Madrid, Spain, 2021: 1–6. doi: 10.1109/GLOBECOM46510.2021.9685067.
[11]	LI Haozhe, LIAO Yilin, WANG Wenhai, et al. A novel time-domain graph tensor attention network for specific emitter identification[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 5501414. doi: 10.1109/TIM.2023.3241976.
[12]	ZHA Xiong, LI Tianyun, YANG Kaiyuan, et al. Open-set radio frequency fingerprint identification via uncertainty awareness[C]. International Conference on Wireless Communications and Signal Processing (WCSP), Hangzhou, China, 2023: 785–790. doi: 10.1109/WCSP58612.2023.10405358.
[13]	YANG Tianwen, ZHAO Jianing, WANG Xin, et al. Deep learning based RFF recognition with differential constellation trace figure towards closed and open set[C]. IEEE/CIC International Conference on Communications in China (ICCC), Foshan, China, 2022: 908–913. doi: 10.1109/ICCC55456.2022.9880623.
[14]	MENG Zepeng, ZHAO Caidan, XIAO Liang, et al. Domain adaptive open-set recognition algorithm based on data augmentation[C]. IEEE/CIC International Conference on Communications in China (ICCC), Hangzhou, China, 2024: 527–532. doi: 10.1109/ICCC62479.2024.10682042.
[15]	LI Kunling, BAO Jiazhong, XIE Xin, et al. Receiver-agnostic radio frequency fingerprint identification for zero-trust wireless networks[J]. IEEE Journal on Selected Areas in Communications, 2025, 43(6): 1981–1997. doi: 10.1109/JSAC.2025.3560002.
[16]	XIE Wei, WANG Hongjun, SHEN Zhexian, et al. A novel radio frequency fingerprint identification scheme for few-shot open-set recognition[J]. IEEE Internet of Things Journal, 2025, 12(13): 25691–25706. doi: 10.1109/JIOT.2025.3559183.
[17]	SHI Feng, WAN Hong, FENG Ziqin, et al. Enhanced radio frequency fingerprint identification using length-robust representation and incremental learning[J]. IEEE Internet of Things Journal, 2025, 12(10): 14709–14719. doi: 10.1109/JIOT.2025.3526579.
[18]	闫高丽, 付雪, 王禹, 等. 面向射频指纹信号分析与智能识别的研究综述[J]. 南通大学学报: 自然科学版, 2025, 24(2): 1–21. YAN Gaoli, FU Xue, WANG Yu, et al. A survey on radio frequency fingerprint signal analysis and intelligent identification[J]. Journal of Nantong University: Natural Science Edition, 2025, 24(2): 1–21.
[19]	SZEGEDY C, LIU Wei, JIA Yangqing, et al. Going deeper with convolutions[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, USA, 2015: 1–9. doi: 10.1109/CVPR.2015.7298594.
[20]	XIE Renjie, XU Wei, CHEN Yanzhi, et al. A generalizable model-and-data driven approach for open-set RFF authentication[J]. IEEE Transactions on Information Forensics and Security, 2021, 16: 4435–4450. doi: 10.1109/TIFS.2021.3106166.
[21]	SHEN Guanxiong, ZHANG Junqing, MARSHALL A, et al. Towards scalable and channel-robust radio frequency fingerprint identification for LoRa[J]. IEEE Transactions on Information Forensics and Security, 2022, 17: 774–787. doi: 10.1109/TIFS.2022.3152404.
[22]	CAI Zhuoran, LIU Zhiyuan, and KOU Liang. Reliable UAV monitoring system using deep learning approaches[J]. IEEE Transactions on Reliability, 2022, 71(2): 973–983. doi: 10.1109/TR.2021.3119068.
[23]	HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016: 770–778. doi: 10.1109/CVPR.2016.90.
[24]	WANG Yinglin, CAO Chunjie, LI Yifan, et al. Radiofrequency fingerprint feature extraction and recognition using a coordinate attention-guided deep residual shrinkage network[C]. International Conference on Networking and Network Applications (NaNA), Qingdao, China, 2023: 551–557. doi: 10.1109/NaNA60121.2023.00097.
[25]	VAN DER MAATEN L and HINTON G. Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(86): 2579–2605.