基于时频循环平稳特征的通信辐射源个体识别

doi:10.12305/j.issn.1001-506X.2026.03.33

摘要/Abstract

摘要：

针对通信辐射源个体识别面临信号特征提取受噪声干扰的问题，构建测量信号数学模型，针对船舶自动识别系统辐射源信号研究。首先，运用短时傅里叶变换推导信号周期平稳特征的时频能量谱，分析其统计量与循环平稳特征关系；其次，提出构建深度学习训练数据集的方法，通过外场实测，表明了循环平稳特征的时频能量谱的差异性、稳定性以及抑制噪声干扰的有效性；最后，利用不同网络和时频特征对训练和测试样本进行比较实验，验证了基于循环平稳特征的累积时频能量谱方法对通信辐射源个体识别准确率的提升效果。针对船舶自动识别系统信号，在10种典型网络模型下的平均Top-1识别准确率为75.92%，相对传统的非累积识别方法性能提高约25%。该方法能有效应对不同时空场景下的噪声干扰，为非合作条件下通信辐射源个体识别提供了一种新的解决方案。

关键词: 辐射源个体识别, 时频能量谱, 短时傅里叶变换, 循环平稳

Abstract:

Addressing the issue of signal feature extraction being affected by noise interference in the individual identification of communication emitters, a mathematical model for measuring signals is constructed, and research is conducted on the emitter signals of the automatic identification system for ships. Firstly, the time-frequency energy spectrum of the signal’s cyclostationary features is derived using the short-time Fourier transform, and the relationship between its statistical quantities and cyclostationary features is analyzed. Secondly, a method for constructing a deep learning training dataset is proposed, through field measurements, the differences and stability of the time-frequency energy spectrum of cyclostationary features, as well as its effectiveness in suppressing noise interference are verified. Finally, comparative experiments are conducted on the training and test samples by using different networks and time-frequency features, which validates that the cumulative time-frequency energy spectrum method based on cyclostationary features can improve the accuracy of communication emitter individual identification. For automatic identification system signals for ships, the average Top-1 identification accuracy across 10 typical network models reaches 75.92%, which is approximately a 25% performance improvement compared to traditional non-cumulative identification methods. The method can effectively suppress noise interference in various spatiotemporal scenarios, providing a novel solution for communication emitter identification under non-cooperative conditions.

Key words: emitter individual identification, time-frequency spectrum, short-time Fourier transform, cyclostationary

中图分类号:

TN 911

黄宇, 张鑫, 田威, 范崧伟, 余立志. 基于时频循环平稳特征的通信辐射源个体识别[J]. 系统工程与电子技术, 2026, 48(3): 1091-1101.

Yu HUANG, Xin ZHANG, Wei TIAN, Songwei FAN, Lizhi YU. Individual identification of communication emitter based on time-frequency cyclostationary features[J]. Systems Engineering and Electronics, 2026, 48(3): 1091-1101.

图/表 14

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

表1

表2

图11

图12

参考文献 33

1	闫文君, 段可欣, 凌青, 等. 数据恶劣条件下的辐射源个体识别方法综述[J]. 海军航空大学学报, 2024, 39 (5): 523- 534.
	YAN W J, DUAN K X, LING Q, et al. Overview of individual identification methods for radiation sources under harsh data conditions[J]. Journal of Naval Aviation University, 2024, 39 (5): 523- 534.
2	LI Q L, WANG F S, ZHANG Y B, et al. Novel micro-UAV identification approach assisted by combined RF fingerprint[J]. IEEE Sensors Journal, 2024, 24 (16): 26802- 26813. doi: 10.1109/JSEN.2024.3411106
3	LU Q, YANG Z K, ZHANG H L, et al. MRFE: a deep-learning-based multidimensional radio frequency fingerprinting enhancement approach for IoT device identification[J]. IEEE Internet of Things Journal, 2024, 11 (18): 30442- 30454. doi: 10.1109/JIOT.2024.3414195
4	LIN Y, ZHA H R, TU Y, et al. GLR-SEI: green and low resource specific emitter identification based on complex networks and fisher pruning[J]. IEEE Trans. on Emerging Topics in Computational Intelligence, 2024, 8 (5): 3239- 3250.
5	YAO H M, ZHANG H H, JIANG L J, et al. Enhanced deep learning approach for electromagnetic forward modeling of dielectric target within the wide frequency band using deep residual convolutional neural network[J]. IEEE Antennas and Wireless Propagation Letters, 2024, 23 (6): 1884- 1888. doi: 10.1109/LAWP.2024.3372437
6	ZHANG J N, LIU Y S, DING G R, et al. Adaptive decomposition and extraction network of individual fingerprint features for specific emitter identification[J]. IEEE Trans. on Information Forensics and Security, 2024, 19, 8515- 8528. doi: 10.1109/TIFS.2024.3427361
7	张顺生, 丁宦城, 王文钦. 面向辐射源识别的多尺度特征提取与特征选择网络[J]. 国防科技大学学报, 2024, 46 (6): 141- 148. doi: 10.11887/j.cn.202406015
	ZHANG S S, DING H C, WANG W Q. Multi-scale feature extraction and feature selection network for radiation source identification[J]. Journal of National University of Defense Technology, 2024, 46 (6): 141- 148. doi: 10.11887/j.cn.202406015
8	PENG Y, LIU P F, ZHANG Q Y, et al. Dynamic adaptation RFF identification method leveraging cognitive representation learning[J]. IEEE Trans. on Information Forensics and Security, 2024, 19, 7939- 7951. doi: 10.1109/TIFS.2024.3451710
9	CHAI Z, PENG X Y, HUANG X R, et al. Channel-robust RF fingerprint identification using multi-task learning and receiver collaboration[J]. IEEE Signal Processing Letters, 2024, 31, 2510- 2514. doi: 10.1109/LSP.2024.3460654
10	ZHANG T T, XU D Y, REN P Y, et al. DFLNet: deep federated learning network with privacy preserving for vehicular LoRa nodes fingerprinting[J]. IEEE Trans. on Vehicular Technology, 2024, 73 (2): 2901- 2905. doi: 10.1109/TVT.2023.3316639
11	张多纳, 赵宏佳, 鲁远耀, 等. 融入注意力机制的小样本遥感图像场景分类[J]. 计算机工程与应用, 2024, 60 (4): 173- 182.
	ZHANG D N, ZHAO H J, LU Y Y, et al. Few-shot scene classification with attention mechanism in remote sensing[J]. Computer Engineering and Applications, 2024, 60 (4): 173- 182.
12	FAN R, SI C K, HAN Y, et al. RFFsNet-SEI: a multidimensional balanced-RFFs deep neural network framework for specific emitter identification[J]. Journal of Systems Engineering and Electronics, 2024, 35 (3): 558- 574. doi: 10.23919/JSEE.2023.000069
13	ABDURRAHMAN E, BECHIR H. Distinguishable IQ feature representation for domain-adaptation learning of WiFi device fingerprints[J]. IEEE Trans. on Machine Learning in Communications and Networking, 2024, 2, 1404- 1423. doi: 10.1109/TMLCN.2024.3446743
14	谢存祥, 张立民, 钟兆根. 基于Hilbert-Huang变换与对抗训练的特定辐射源识别[J]. 系统工程与电子技术, 2021, 43 (12): 3478- 3487. doi: 10.12305/j.issn.1001-506X.2021.12.08
	XIE C X, ZHANG L M, ZHONG Z G. Specific emitter identification based on Hilbert-Huang transform and adversarial training[J]. Systems Engineering and Electronics, 2021, 43 (12): 3478- 3487. doi: 10.12305/j.issn.1001-506X.2021.12.08
15	LI B, CETIN E. A graphical deep learning approach to RF fingerprinting in the time-frequency domain[J]. IEEE Sensors Journal, 2024, 24 (10): 16984- 16990. doi: 10.1109/JSEN.2024.3383437
16	SHEN G X, ZHANG J Q, ALAN M, et al. Towards scalable and channel-robust radio frequency fingerprint identification for LoRa[J]. IEEE Trans. on Information Forensics and Security, 2022, 17, 774- 787. doi: 10.1109/TIFS.2022.3152404
17	王磊, 张志勇, 胥辉旗, 等. 基于VMD分解和多域联合分布的雷达辐射源识别[J]. 系统工程与电子技术, 2023, 45 (8): 2479- 2488.
	WANG L, ZHANG Z Y, XU H Q, et al. Radar emitter identification based on VMD and multi-domain jointed distribution[J]. Systems Engineering and Electronics, 2023, 45 (8): 2479- 2488.
18	JIANG Q, SHA J. RF fingerprinting identification in low SNR scenarios for automatic identification system[J]. IEEE Trans. on Wireless Communications, 2024, 23 (3): 2070- 2081. doi: 10.1109/TWC.2023.3294988
19	FU H, SUN Y F, PENG L N, et al. Channel-resilient RF fingerprint identification based on nonlinear features with memory effect[J]. IEEE Communications Letters, 2024, 28 (4): 798- 802. doi: 10.1109/LCOMM.2024.3364494
20	AMEYA S R, JACOB D R, LENNART M H, et al. Robust measurements for RF fingerprinting with constellation patterns of radiated waveforms [C]//Proc. of the IEEE Physical Assurance and Inspection of Electronics, 2023.
21	YIN P C, PENG L N, SHEN G X, et al. Multi-channel CNN-based open-set RF fingerprint identification for LTE devices[J]. IEEE Trans. on Communications and Networking, 2024, 10 (5): 1788- 1800. doi: 10.1109/TCCN.2024.3391293
22	PENG L N, WU Z N, ZHANG J Q, et al. Hybrid RFF identification for LTE using wavelet coefficient graph and differential spectrum[J]. IEEE Trans. on Vehicular Technology, 2024, 73 (8): 11621- 11636. doi: 10.1109/TVT.2024.3380671
23	AMANI AS, FRANCESCO R, SALVATORE DO, et al. Exposing the fingerprint: dissecting the impact of the wireless channel on radio fingerprinting [C]//Proc. of the IEEE Conference on Computer Communications, 2020: 646−655.
24	QI X Y, HU A Q, CHEN T S. Lightweight radio frequency fingerprint identification scheme for V2X based on temporal correlation[J]. IEEE Trans. on Information Forensics and Security, 2024, 19, 1056- 1070. doi: 10.1109/TIFS.2023.3329683
25	PENG L N, PENG H C, FU H. Channel-robust radio frequency fingerprint identification for cellular uplink LTE devices[J]. IEEE Internet of Things Journal, 2024, 11 (10): 17154- 17169. doi: 10.1109/JIOT.2024.3358904
26	HUANG L, BI S Z, ZHANG Y J. Deep reinforcement learning for online computation offloading in wireless powered mobile-edge computing networks[J]. IEEE Trans. on Mobile Computing, 2020, 19 (11): 2581- 2593. doi: 10.1109/TMC.2019.2928811
27	BRAEDEN P M, LAUREN J W, ALAN J M. Sensitivity analysis of RFML applications[J]. IEEE Access, 2024, 12, 80327- 80344. doi: 10.1109/ACCESS.2024.3409471
28	GAO Y, WU Q Q, CHEN W, et al. Exploiting intelligent reflecting surfaces for interference channels with SWIPT[J]. IEEE Trans. on Wireless Communications, 2024, 23 (5): 4442- 4458. doi: 10.1109/TWC.2023.3318795
29	CAI Z X, WANG Y, JIANG Q, et al. Toward intelligent lightweight and efficient UAV identification with RF fingerprinting[J]. IEEE Internet of Things Journal, 2024, 11 (15): 26329- 26339. doi: 10.1109/JIOT.2024.3395466
30	TANG P, DING G R, XU Y T, et al. Causal learning for robust specific emitter identification over unknown channel statistics[J]. IEEE Trans. on Information Forensics and Security, 2024, 19, 5316- 5329. doi: 10.1109/TIFS.2024.3393237
31	BAI N, INWHEE J. Deep learning methods with the improved attention for explainable image recognition[J]. IEEE Access, 2024, 12, 70559- 70567. doi: 10.1109/ACCESS.2024.3397323
32	ZHOU Q, GAO J Y, YUAN Y, et al. Single-stream extractor network with contrastive pre-training for remote-sensing change captioning[J]. IEEE Trans. on Geoscience and Remote Sensing, 2024, 62, 5624514.
33	CHEN J Y, GENG Y X, CHEN Z, et al. Zero-shot and few-shot learning with knowledge graphs: a comprehensive survey[J]. Proceedings of the IEEE, 2023, 111 (6): 653- 685. doi: 10.1109/JPROC.2023.3279374

方法	平均	Google Net	ResNet 50	DarkNet 53	VGG 16	VGG 19	Inception Resnetv2	Incept ionv3	NasNet Mobile	ResNet101	DenseNet201
HOSA	0.4270	0.3787	0.4012	0.5418	0.3942	0.3132	0.4576	0.5062	0.4310	0.4238	0.4220
WVD	0.5085	0.5104	0.4962	0.5601	0.4635	0.4909	0.4799	0.5165	0.5189	0.5242	0.5243
STFT	0.4702	0.3642	0.5062	0.5643	0.3752	0.4182	0.4606	0.5531	0.4755	0.4720	0.5127
STFT累积	0.7592	0.6024	0.6401	0.7079	0.7297	0.7273	0.7120	0.8235	0.8391	0.8823	0.9279

参数	Google Net	ResNet 50	DarkNet 53	VGG 16	VGG 19	Inception Resnetv2	Incept ionv3	NasNet Mobile	ResNet101	DenseNet201
深度	22	50	53	16	19	164	48	—	101	201
大小/MB	27	96	155	515	535	209	89	20	167	77
参数(×10⁵)	7	25.6	41.6	138	144	55.9	23.9	5.3	44.6	20
输入大小	224×224×3		256×256×3	224×224×3		299×299×3		224×224×3

[1]	钟琦, 魏绍仁, 贺志毅, 刘建欣. 基于时频域行聚类的抗间歇采样转发干扰方法[J]. 系统工程与电子技术, 2026, 48(1): 106-118.
[2]	陈宇鹏, 黄科举, 刘辉, 邝龙坤, 杨俊安. 基于信号递归图和卷积宽度学习的小样本辐射源个体识别方法[J]. 系统工程与电子技术, 2025, 47(9): 3086-3092.
[3]	赵正义, 南普龙. 基于短时傅里叶变换的高效空频自适应抗干扰方法[J]. 系统工程与电子技术, 2025, 47(6): 1778-1785.
[4]	刘佳楠, 宋志群, 李勇, 刘丽哲, 夏金涛, 王斌, 李行健, 汪畅. 基于深度学习的复合干扰信号识别[J]. 系统工程与电子技术, 2025, 47(10): 3512-3519.
[5]	杨磊, 孙卫天, 毛欣瑶, 夏亚波, 桑婧隺. 雷达鸟类目标微多普勒贝叶斯增强算法[J]. 系统工程与电子技术, 2024, 46(2): 505-516.
[6]	王春升, 王永民, 许华, 朱华丽. 基于残差原型网络的辐射源个体识别[J]. 系统工程与电子技术, 2023, 45(7): 2249-2258.
[7]	吕勤哲, 全英汇, 沙明辉, 董淑仙, 邢孟道. 基于集成深度学习的有源干扰智能分类[J]. 系统工程与电子技术, 2022, 44(12): 3595-3602.
[8]	陈晓威, 杨文革. 基于时变白化滤波器的DSSS系统干扰盲检测[J]. 系统工程与电子技术, 2021, 43(7): 1981-1988.
[9]	王曼颖, 龚晓峰, 雒瑞森, 边彤, 王之远. 基于自适应形态学的跳频信号参数联合盲估计[J]. 系统工程与电子技术, 2021, 43(5): 1398-1405.
[10]	陈浩, 杨俊安, 刘辉. 基于深度残差适配网络的通信辐射源个体识别[J]. 系统工程与电子技术, 2021, 43(3): 603-609.
[11]	刘海涛, 李广雪, 孙海霞, 王磊. L-DACS1系统干扰抑制与循环自适应波束形成方法[J]. 系统工程与电子技术, 2020, 42(8): 1850-1856.
[12]	吴子龙, 陈红, 雷迎科, 李昕, 熊颢. 基于堆栈式LSTM网络的通信辐射源个体识别[J]. 系统工程与电子技术, 2020, 42(12): 2915-2923.
[13]	方章闻, 张金艺, 李科, 姜玉稀. 小样本条件下的通信辐射源半监督特征提取[J]. 系统工程与电子技术, 2020, 42(10): 2381-2389.
[14]	刘明骞, 颜志文, 张俊林. 空中目标辐射源的个体识别方法[J]. 系统工程与电子技术, 2019, 41(11): 2408-2415.
[15]	王桂良, 黄渊凌. 多普勒及多普勒变化率对辐射源个体识别性能的影响分析[J]. 系统工程与电子技术, 2017, 39(12): 2671-2676.