基于SVM与K-means的多层级雷达信号开集识别

翁雪慧; 汪晓锋; 应鹏; 刘重安; 周芳; 全大英

doi:10.13700/j.bh.1001-5965.2024.0369

基于SVM与K-means的多层级雷达信号开集识别

doi: 10.13700/j.bh.1001-5965.2024.0369

中国计量大学信息工程学院，杭州 310018

基金项目:

国家自然科学基金(62261014)

详细信息

通讯作者:
E-mail：xfwang@cjlu.edu.cn

中图分类号: V221⁺.3；TB553
计量
- 文章访问数: 253
- HTML全文浏览量: 81
- PDF下载量: 32
- 被引次数: 0
出版历程
- 收稿日期: 2024-05-30
- 录用日期: 2025-03-21
- 网络出版日期: 2025-04-14
- 整期出版日期: 2026-07-31

Multi-level radar signal open-set recognition based on SVM and K-means

College of Information Engineering，China Jiliang University，Hangzhou 310018，China

Funds:

National Natural Science Foundation of China (62261014)

More Information

Corresponding author: E-mail：xfwang@cjlu.edu.cn

摘要

摘要:
针对常见雷达信号调制识别方法在真实场景中难以有效识别未知调制信号的问题，提出一种基于支持向量机(SVM)预训练和K-means聚类的多层级雷达信号开集识别方法。通过将雷达信号进行多重同步压缩(MSST)时频变换后，利用小波变换(DWT)对预处理后的时频图像进行特征提取。在训练阶段，利用已知雷达信号数据来训练一个SVM分类器。在测试阶段，利用该分类器来划分已知调制方式和未知调制方式，并完成已知雷达信号的闭集识别。对未知类信号则利用K-means进行聚类分析，将未知调制方式进一步划分到不同类别，以扩展雷达信号调制方式的识别范围。实验结果表明：所提方法在信噪比(SNR)为−4 dB时对已知信号和未知信号的识别准确率均达到90%以上，实现了对未知调制方式的有效识别。
- 雷达信号识别 /
- 多重同步压缩 /
- 小波变换 /
- 开集识别 /
- K-means聚类
Abstract:
In order to address the issue that traditional radar signal recognition techniques have trouble successfully identifying unknown modulated signals in practical situations, this paper suggests a multi-level radar signal open-set identification technique based on K-means clustering and Support Vector Machine (SVM) pre-training. After performing the multisynchro squeezing transform (MSST) on radar signals, the discrete wavelet transform (DWT) is employed to extract features from the preprocessed time-frequency images. An SVM classifier is trained using known radar signal data during the training phase. The classifier is used to distinguish between known and unknown modulation types during the testing phase in order to achieve open-set radar signal identification. Subsequently, K-means cluster analysis is applied to unknown radar signals, further classifying unknown modulation modes into different clusters, thereby expanding the recognition scope of radar signal modulation types. Experimental results demonstrate that the proposed method achieves a recognition accuracy of over 90% for both known and unknown signals at a signal-to-noise ratio (SNR) of −4 dB, effectively recognizing unknown modulation types.
- radar signal recognition /
- multiple synchronous compression /
- wavelet transform /
- open-set recognition /
- K-means clustering

HTML全文

图 1 本文方法整体流程图

Figure 1. Overall process of the proposed method

下载: 全尺寸图片幻灯片

图 2 多种典型雷达信号MSST时频图

Figure 2. MSST time-frequency images of multiple typical radar signals

下载: 全尺寸图片幻灯片

图 3 时频图像预处理

Figure 3. Time-frequency image preprocessing

下载: 全尺寸图片幻灯片

图 4 小波变换分解过程

Figure 4. Decomposition process of wavelet transform

下载: 全尺寸图片幻灯片

图 5 4种方法的平均识别准确率

Figure 5. The average recognition accuracy of four method

下载: 全尺寸图片幻灯片

图 6 6类已知信号的识别准确率

Figure 6. Recognition accuracy of 6 types of known signals

下载: 全尺寸图片幻灯片

图 7 开集识别准确率

Figure 7. Open-set recognition accuracy

下载: 全尺寸图片幻灯片

图 8 −2 dB下的识别混淆矩阵

Figure 8. Confusion matrix of identification at −2 dB

下载: 全尺寸图片幻灯片

图 9 未知雷达信号聚类准确率

Figure 9. Clustering accuracy of unknown radar signals

下载: 全尺寸图片幻灯片

图 10 不同置信度下的准确率

Figure 10. Accuracy under different confidence levels

下载: 全尺寸图片幻灯片

表 1 雷达信号参数

Table 1. Radar signal parameters

雷达信号参数	取值
CW载频/MHz	20
QPSK载频/MHz	20
LFM载频/MHz	20
NLFM载频/MHz	20
FSK/BPSK载频/MHz	20
LFM/BPSK载频/MHz	20
LFM/FSK载频/MHz	20
2FSK两频率点选取范围/MHz	10～30
BPSK载频/MHz	20
T1载频/MHz	20
T1序列段数	｛4,5,6｝
注：采样频率为200 MHz，脉宽为3.5 μm。

下载: 导出CSV

参考文献(36)

[1]	李猛. 现代雷达信号处理的技术和发展趋势[J]. 电子通信与计算机科学, 2023, 5(2): 7-9. Li M. Technology and development trend of modern radar signal processing[J]. Electronic Communication and Computer Science, 2023, 5(2): 7-9(in Chinese).
[2]	张怀国. 雷达信号识别关键技术研究[J]. 电子技术与软件工程, 2020, 9(10): 85-87. Zhang H G. Research on key technologies of radar signal recognition[J]. Electronic Technology & Software Engineering, 2020, 9(10): 85-87(in Chinese).
[3]	熊孝华, 杨安会. 现代雷达信号处理及发展趋势研究[J]. 中国高新技术企业, 2011(13): 23-24. Xiong X H, Yang A H. Research on modern radar signal processing and its development trend[J]. China High-Tech Enterprises, 2011(13): 23-24(in Chinese).
[4]	蒋千军, 张明安, 卢敏, 等. 现代雷达信号处理的技术和发展趋势探讨[J]. 电子通信与计算机科学, 2023, 5(3): 131-133. Jiang Q J, Zhang M A, Lu M, et al. Discussion on the technology and development trend of modern radar signal processing[J]. Electronic Communication and Computer Science, 2023, 5(3): 131-133(in Chinese).
[5]	王海英, 张群英, 成文海, 等. LPI雷达信号调制识别及参数估计研究进展[J]. 系统工程与电子技术, 2024, 46(6): 1908-1924. Wang H Y, Zhang Q Y, Cheng W H, et al. Research progress on LPI radar signal modulation recognition and parameter estimation[J]. Systems Engineering and Electronics, 2024, 46(6): 1908-1924(in Chinese).
[6]	管飞, 宁云晖, 李超, 等. 雷达侦察接收机电磁环境评估和分级方法[J]. 舰船电子工程, 2024, 44(1): 209-213. Guan F, Ning Y H, Li C, et al. Evaluation and classification of electromagnetic environment of radar reconnaissance receiver[J]. Ship Electronic Engineering, 2024, 44(1): 209-213(in Chinese).
[7]	刘歌, 张国毅, 于岩. 基于随机森林的雷达信号脉内调制识别[J]. 电信科学, 2016, 32(5): 69-76. Liu G, Zhang G Y, Yu Y. Intra-pulse modulation recognition of radar signal based on random forest[J]. Telecommunications Science, 2016, 32(5): 69-76(in Chinese).
[8]	杨志远, 张玉玺, 孙进平, 等. 基于CNN的雷达辐射源信号开集识别方法[C]//第十四届全国DSP应用技术学术会议论文集. 北京: 中国电子学会数字信号处理专家委员会, 2022: 113-117. Yang Z Y, Zhang Y X, Sun J P, et al. Open set recognition method of radar emitter signal based on CNN[C]//Proceedings of the 14th National DSP Application Technology Academic Conference. Beijing: Expert Committee on Digital Signal Processing of the Chinese Electronics Society, 2022: 113-117(in Chinese).
[9]	张静, 于蕾, 侯长波, 等. 基于时频特征的跳频信号调制识别[J]. 太赫兹科学与电子信息学报, 2022, 20(1): 40-46. Zhang J, Yu L, Hou C B, et al. Frequency-hopping signal modulation recognition based on time-frequency features[J]. Journal of Terahertz Science and Electronic Information Technology, 2022, 20(1): 40-46(in Chinese).
[10]	李世通, 全大英, 唐泽雨, 等. 基于时频图像和高次频谱特征的雷达信号识别[J]. 电信科学, 2022, 38(2): 84-91. Li S T, Quan D Y, Tang Z Y, et al. Time-frequency image and high-order spectrum characteristics based radar signal recognition[J]. Telecommunications Science, 2022, 38(2): 84-91(in Chinese).
[11]	Qu Z Y, Hou C F, Hou C B, et al. Radar signal intra-pulse modulation recognition based on convolutional neural network and deep Q-learning network[J]. IEEE Access, 2020, 8: 49125-49136.
[12]	姬丽彬, 朱岩, 崔天舒, 等. 基于时频重排算法的LPI雷达信号识别[J]. 北京航空航天大学学报, 2025, 51(4): 1324-1331. Ji L B, Zhu Y, Cui T S, et al. LPI radar signal recognition based on time-frequency reassignment algorithm[J]. Journal of Beijing University of Aeronautics and Astronautics, 2025, 51(4): 1324-1331(in Chinese).
[13]	郭立民, 陈鑫, 陈涛. 基于AlexNet模型的雷达信号调制类型识别[J]. 吉林大学学报(工学版), 2019, 49(3): 1000-1008. Guo L M, Chen X, Chen T. Radar signal modulation type recognition based on AlexNet model[J]. Journal of Jilin University (Engineering and Technology Edition), 2019, 49(3): 1000-1008(in Chinese).
[14]	谢存祥, 张立民, 钟兆根. 基于时频特征提取和残差神经网络的雷达信号识别[J]. 系统工程与电子技术, 2021, 43(4): 917-926. Xie C X, Zhang L M, Zhong Z G. Radar signal recognition based on time-frequency feature extraction and residual neural network[J]. Systems Engineering and Electronics, 2021, 43(4): 917-926(in Chinese).
[15]	Li P. Research on radar signal recognition based on automatic machine learning[J]. Neural Computing and Applications, 2020, 32(7): 1959-1969.
[16]	韩啸, 陈世文, 陈蒙, 等. 基于互易点学习的LPI信号开集识别[J]. 系统工程与电子技术, 2022, 44(9): 2752-2759. Han X, Chen S W, Chen M, et al. Open-set recognition of LPI radar signal based on reciprocal point learning[J]. Systems Engineering and Electronics, 2022, 44(9): 2752-2759(in Chinese).
[17]	郝云飞, 刘章孟, 郭福成, 等. 基于生成对抗网络的信号调制方式的开集识别[J]. 系统工程与电子技术, 2019, 41(11): 2619-2624. Hao Y F, Liu Z M, Guo F C, et al. Open-set recognition of signal modulation based on generative adversarial networks[J]. Systems Engineering and Electronics, 2019, 41(11): 2619-2624(in Chinese).
[18]	曹鹏宇, 杨承志, 石礼盟, 等. 基于PSO-DBSCAN和SCGAN的未知雷达信号处理方法[J]. 系统工程与电子技术, 2022, 44(4): 1158-1165. Cao P Y, Yang C Z, Shi L M, et al. Unknown radar signal processing based on PSO-DBSCAN and SCGAN[J]. Systems Engineering and Electronics, 2022, 44(4): 1158-1165(in Chinese).
[19]	秦博伟, 蒋磊, 许华, 等. 基于RE-GAN的调制信号开集识别算法[J]. 系统工程与电子技术, 2023, 45(10): 3321-3328. Qin B W, Jiang L, Xu H, et al. Open-set recognition algorithm for modulation signal based on RE-GAN[J]. Systems Engineering and Electronics, 2023, 45(10): 3321-3328(in Chinese).
[20]	Shi Y L, Quan D Y, Zhang Z, et al. Radar signal recognition based on MSST and dual channel feature extraction[C]//Proceedings of the IEEE 18th Conference on Industrial Electronics and Applications. Piscataway: IEEE Press, 2023: 358-364.
[21]	Guo S Z, Akhtar S, Mella A. A method for radar model identification using time-domain transient signals[J]. IEEE Transactions on Aerospace and Electronic Systems, 2021, 57(5): 3132-3149.
[22]	Li R N, Yang Z W, He S, et al. Ship target micro-Doppler feature extracted by multi-synchrosqueezing transform and inverse radon transform[C]//Proceedings of the CIE International Conference on Radar. Piscataway: IEEE Press, 2023: 1545-1549.
[23]	Li X Y, Zhao H M, Yu L, et al. Feature extraction using parameterized multisynchrosqueezing transform[J]. IEEE Sensors Journal, 2022, 22(14): 14263-14272.
[24]	Syed S H, Muralidharan V. Feature extraction using Discrete Wavelet Transform for fault classification of planetary gearbox–A comparative study[J]. Applied Acoustics, 2022, 188: 108572.
[25]	Kirti, Rajpal N, Yadav J. A novel DWT and deep learning based feature extraction technique for plant disease identification[C]//Proceedings of the Second Doctoral Symposium on Computational Intelligence. Singapore: Springer, 2022: 355-367.
[26]	Xiong L Z, Zhong X W, Yang C N. DWT-SISA: a secure and effective discrete wavelet transform-based secret image sharing with authentication[J]. Signal Processing, 2020, 173: 107571.
[27]	Kumar A, Tomar H, Mehla V K, et al. Stationary wavelet transform based ECG signal denoising method[J]. ISA Transactions, 2021, 114: 251-262.
[28]	Barjasteh A, Ghafouri S H, Hashemi M. A hybrid model based on discrete wavelet transform (DWT) and bidirectional recurrent neural networks for wind speed prediction[J]. Engineering Applications of Artificial Intelligence, 2024, 127: 107340.
[29]	Omidvar M, Zahedi A, Bakhshi H. EEG signal processing for epilepsy seizure detection using 5-level Db4 discrete wavelet transform, GA-based feature selection and ANN/SVM classifiers[J]. Journal of Ambient Intelligence and Humanized Computing, 2021, 12(11): 10395-10403.
[30]	Jiang J, Ma K, Wen M, et al. An audio fingerprint algorithm based on statistical characteristics of db4 wavelet[J]. Journal of Information and Computational Science, 2011, 8(14): 3027-3034.
[31]	Chandra M A, Bedi S S. Survey on SVM and their application in imageclassification[J]. International Journal of Information Technology, 2021, 13(5): 1-11.
[32]	Sinaga K P, Yang M S. Unsupervised K-means clustering algorithm[J]. IEEE Access, 2020, 8: 80716-80727.
[33]	Ghazal T M, Zahid Hussain M, Said R A, et al. Performances of K-means clustering algorithm with different distance metrics[J]. Intelligent Automation & Soft Computing, 2021, 29(3): 735-742.
[34]	Ikotun A M, Ezugwu A E, Abualigah L, et al. K-means clustering algorithms: a comprehensive review, variants analysis, and advances in the era of big data[J]. Information Sciences, 2023, 622: 178-210.
[35]	全大英, 唐泽雨, 陈赟, 等. 基于MSST及HOG特征提取的雷达辐射源信号识别[J]. 北京航空航天大学学报, 2023, 49(3): 538-547. Quan D Y, Tang Z Y, Chen Y, et al. Radar emitter signal recognition based on MSST and HOG feature extraction[J]. Journal of Beijing University of Aeronautics and Astronautics, 2023, 49(3): 538-547(in Chinese).
[36]	王功明. 雷达信号脉内特征分析与识别关键技术研究[D]. 郑州: 战略支援部队信息工程大学, 2019. Wang G M. Research on key technology of radar signal intra-pulse features analysis and recognition[D]. Zhengzhou: Information Engineering University, 2019(in Chinese).