基于单通道多尺度图神经网络的自动调制识别

国强; 聂孟允; 戚连刚; Kaliuzhnyi Mykola

doi:10.11999/JEIT220840

基于单通道多尺度图神经网络的自动调制识别

doi: 10.11999/JEIT220840

国强^{1, 2},
聂孟允^{1, 2},
戚连刚^{1, 2, ,},
Kaliuzhnyi Mykola^{1, 3}

1.
哈尔滨工程大学信息与通信工程学院哈尔滨 150001
2.
先进船舶通信与信息技术工业和信息化部重点实验室哈尔滨 150001
3.
哈尔科夫国立无线电电子大学哈尔科夫 61166

基金项目: 国家重点研发计划(2018YFE0206500)，国家自然科学基金(62071140)，中央高校基本科研业务费专项资金(3072022QBZ0801)

详细信息

作者简介:
国强：男，教授，研究方向为电子对抗、智能信号处理与识别

聂孟允：男，硕士生，研究方向为基于深度学习的信号识别

戚连刚：男，讲师，研究方向为自适应干扰抑制、卫星导航信号处理

Kaliuzhnyi Mykola：男，教授，研究方向为电子对抗、电磁目标识别

通讯作者:
戚连刚　qiliangang@hrbeu.edu.cn

中图分类号: TN911
计量
- 文章访问数: 809
- HTML全文浏览量: 381
- PDF下载量: 200
- 被引次数: 0
出版历程
- 收稿日期: 2022-06-24
- 修回日期: 2022-11-08
- 网络出版日期: 2022-11-10
- 刊出日期: 2023-05-10

Automatic Modulation Recognition Based on Single-channel Multi-scale Graph Neural Network

GUO Qiang^{1, 2},
NIE Mengyun^{1, 2},
QI Liangang^{1, 2
, ,},
Kaliuzhnyi Mykola^{1, 3}

1.
College of Information and Communication Engineering, Harbin Engineering University, Harbin 150001, China
2.
Key Laboratory of Advanced Marine Communication and Information Technology Ministry of Industry and Information Technology, Harbin 150001, China
3.
Kharkiv National University of Radio Electronics, Kharkiv 61166, Ukraine

Funds: The National Key R & D Plan(2018YFE0206500), The National Natural Science Foundation of China (62071140), The Fundamental Research Funds for Central Universities (3072022QBZ0801)

摘要

摘要: 针对自适应可见性图(AVG)算法复杂度过高且精度提升不明显的缺点，该文提出一种基于单通道多尺度图神经网络(SMGNN)的自动调制识别(AMR)框架，并对框架各个部分进行了可解释性研究。首先利用多层感知机和1维卷积自适应地实现了单通道信号序列和图之间的映射，有效降低了AVG算法的复杂度；其次，设计了一种多尺度图神经网络，将不同分辨率的特征进行融合，提升了模型识别准确率。实验表明，该文提出的SMGNN算法相比于AVG算法节省了近1/2的参数量，且识别精度得到了较大的提升。
- 自动调制识别 /
- 图神经网络 /
- 自适应可见性图 /
- 特征融合
Abstract: Considering the shortcomings of the Adaptive Visibility Graph (AVG) algorithm being too complex and the accuracy improvement is not significant, an Automatic Modulation Recognition(AMR) framework based on Single-channel Multi-scale Graph Neural Network (SMGNN) is proposed and interpretability studies are conducted on the various parts of the framework. Firstly, the multi-layer perceptron and one-dimensional convolutional adaptive are used to realize the mapping between single-channel signal sequences and graphs, which reduces effectively the complexity of AVG algorithms. Secondly, a multi-scale graph neural network is designed to fuse the features of different resolutions, which improves the accuracy of model recognition. Experiments show that the SMGNN algorithm proposed in this paper saves nearly half of the parameter amount compared with the AVG algorithm, and the recognition accuracy has been greatly improved.
- Automatic Modulation Recognition(AMR) /
- Graph Neural Network(GNN) /
- Adaptive Visibility Graph(AVG) /
- Feature Fusion(FF)

HTML全文

图 1 基于单通道多尺度图神经网络的自动调制识别

下载: 全尺寸图片幻灯片

图 2 通道融合网络及激活函数

下载: 全尺寸图片幻灯片

图 3 自适应图表示网络框架

下载: 全尺寸图片幻灯片

图 4 多尺度图神经网络

下载: 全尺寸图片幻灯片

图 5 矢量调制图

下载: 全尺寸图片幻灯片

图 6 VG算法示意图

下载: 全尺寸图片幻灯片

图 7 基于T-SNE的可视化分析

下载: 全尺寸图片幻灯片

图 8 不同超参数m下的识别率

下载: 全尺寸图片幻灯片

图 9 不同算法的识别精度

下载: 全尺寸图片幻灯片

图 10 不同算法的混淆矩阵分析

下载: 全尺寸图片幻灯片

图 11 消融实验的混淆矩阵分析

下载: 全尺寸图片幻灯片

图 12 单通道可见性图

下载: 全尺寸图片幻灯片

表 1 数据集RadioML2016.10b的参数设置

参数	定义值
采样速率(kHz)	200
信噪比(dB)	–20～18，间隔为2
数据格式	IQ数据格式：2×128
信号数量	1 200 000

下载: 导出CSV

表 2 不同算法对数据集RML2016.10b的识别性能

算法	识别精度(%)	参数量	平均推理时间(s)
CNN2D	55.41	2706954	0.0192
RESNET	63.95	191946	0.0053
LSTM	63.72	197642	0.0064
MCLDNN	64.45	403722	0.0160
AVGNET	64.61	2311984	0.0310
SMGNN	65.36	1246720	0.0176

下载: 导出CSV

表 3 消融实验识别性能

算法	识别精度(%)	参数量	平均推理时间(s)
SMGNNo(单通道单尺度)	64.82	1156352	0.0165
SMGNNs(48维节点特征)	64.81	712976	0.0165
SMGNNs(32维节点特征)	64.67	327712	0.0154
SMGNNs(16维节点特征)	64.48	90928	0.0149

下载: 导出CSV

参考文献(23)

[1]	LIANG Yingchang, TAN Junjie, and NIYATO D. Overview on intelligent wireless communication technology[J]. Journal on Communications, 2020, 41(7): 1–17. doi: 10.11959/j.issn.1000-436x.2020145
[2]	XU J L, SU Wei, and ZHOU Mengchu. Likelihood-Ratio approaches to automatic modulation classification[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part C, 2011, 41(4): 455–469. doi: 10.1109/TSMCC.2010.2076347
[3]	XU J L, SU Wei, and ZHOU Mengchu. Software-Defined radio equipped with rapid modulation recognition[J]. IEEE Transactions on Vehicular Technology, 2010, 59(4): 1659–1667. doi: 10.1109/TVT.2010.2041805
[4]	O’SHEA T J, CORGAN J, and CLANCY T C. Convolutional radio modulation recognition networks[C]. 17th International Conference on Engineering Applications of Neural Networks, Aberdeen, UK, 2016: 213–226.
[5]	WEST N E and O'SHEA T. Deep architectures for modulation recognition[C]. 2017 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN), Baltimore, USA, 2017: 1–6.
[6]	郭业才, 姚文强. 基于信噪比分类网络的调制信号分类识别算法[J]. 电子与信息学报, 2022, 44(10): 3507–3515. doi: 10.11999/JEIT210825 GUO Yecai and YAO Wenqiang. Modulation signal classification and recognition algorithm based on signal to noise ratio classification network[J]. Journal of Electronics &Information Technology, 2022, 44(10): 3507–3515. doi: 10.11999/JEIT210825
[7]	颜康, 金炜东, 黄颖坤, 等. 基于元学习的畸变雷达电磁信号识别[J]. 电子与信息学报, 2022, 44(4): 1351–1357. doi: 10.11999/JEIT210190 YAN Kang, JIN Weidong, HUANG Yingkun, et al. Distorted radar electromagnetic signal recognition based on meta-learning[J]. Journal of Electronics &Information Technology, 2022, 44(4): 1351–1357. doi: 10.11999/JEIT210190
[8]	XUE Fuzhao, SUN Aixin, ZHANG Hao, et al. GDPNet: Refining latent multi-view graph for relation extraction[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2021, 35(16): 14194–14202. doi: 10.1609/aaai.v35i16.17670
[9]	NAN Guoshun, LUO Guoqing, LENG Sicong, et al. Speaker-oriented latent structures for dialogue-based relation extraction[J]. arXiv preprint arXiv: 2109.05182, 2021.
[10]	LACASA L, LUQUE B, BALLESTEROS F, et al. From time series to complex networks: The visibility graph[J]. Proceedings of the National Academy of Sciences of the United States of America, 2008, 105(13): 4972–4975. doi: 10.1073/pnas.0709247105
[11]	LUQUE B, LACASA L, BALLESTEROS F, et al. Horizontal visibility graphs: Exact results for random time series[J]. Physical Review E, 2009, 80(4 Pt 2): 046103.
[12]	QI Xuan, ZHOU Jinchao, QIU Kunfeng, et al. CLPVG: Circular limited penetrable visibility graph as a new network model for time series[J]. Chaos:An Interdisciplinary Journal of Nonlinear Science, 2022, 32(1): 013130. doi: 10.1063/5.0048243
[13]	WAN Tao, JIANG Kaili, TANG Yanli, et al. Automatic LPI radar signal sensing method using visibility graphs[J]. IEEE Access, 2020, 8: 159650–159660. doi: 10.1109/ACCESS.2020.3020336
[14]	LIU Yabo, LIU Yi, and YANG Cheng. Modulation recognition with graph convolutional network[J]. IEEE Wireless Communications Letters, 2020, 9(5): 624–627. doi: 10.1109/LWC.2019.2963828
[15]	QI Xuan, ZHOU Jinchao, QIU Kunfeng, et al. AvgNet: Adaptive visibility graph neural network and its application in modulation classification[J]. IEEE Transactions on Network Science and Engineering, 2022, 9(3): 1516–1526. doi: 10.1109/TNSE.2022.3146836
[16]	YING R, YOU Jiaxuan, MORRIS C, et al. Hierarchical graph representation learning with differentiable pooling[C]. Proceedings of the 32nd International Conference on Neural Information Processing Systems, Montréal, Canada, 2018: 4805–4815.
[17]	VAN DER MAATEN L and HINTON G. Visualizing data using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(2605): 2579–2605.
[18]	樊昌信, 曹丽娜. 通信原理[M]. 7版. 北京: 国防工业出版社, 2012: 176–225. FAN Changxin and CAO Lina. Principles of Communication[M]. 7th ed. Beijing: National Defense Industry Press, 2012: 176–225.
[19]	LIN T Y, DOLLAR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]. The 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, USA, 2017: 936–944.
[20]	O’SHEA T J, ROY T, and CLANCY T C. Over-the-air deep learning based radio signal classification[J]. IEEE Journal of Selected Topics in Signal Processing, 2018, 12(1): 168–179. doi: 10.1109/JSTSP.2018.2797022
[21]	RAJENDRAN S, MEERT W, GIUSTINIANO D, et al. Deep learning models for wireless signal classification with distributed low-cost spectrum sensors[J]. IEEE Transactions on Cognitive Communications and Networking, 2018, 4(3): 433–445. doi: 10.1109/TCCN.2018.2835460
[22]	XU Jialang, LUO Chunbo, PARR G, et al. A spatiotemporal multi-channel learning framework for automatic modulation recognition[J]. IEEE Wireless Communications Letters, 2020, 9(10): 1629–1632. doi: 10.1109/LWC.2020.2999453
[23]	BASTIAN M, HEYMANN S, and JACOMY M. Gephi: An open source software for exploring and manipulating networks[J]. Proceedings of the International AAAI Conference on Web and Social Media, 2009, 3(1): 361–362. doi: 10.1609/icwsm.v3i1.13937