基于深度残差收缩注意力网络的雷达信号识别方法

doi:10.12305/j.issn.1001-506X.2023.03.12

摘要/Abstract

摘要：

针对低信噪比条件下雷达信号识别率低, 以及分类网络不具备识别样本库新添加信号类型的局限, 提出了一种基于深度残差收缩注意力网络的雷达信号识别方法。通过网络将一维雷达信号映射到32维向量空间。网络中的残差连接能有效强化特征的传播能力, 解决网络过深无法训练的问题; 注意力机制的引入, 不仅构建掩码支路充当主干支路的特征选择器, 还能够帮助网络自适应地选择合适的阈值进行软阈值化, 从而减少网络中噪声或者冗余信息的影响, 提高网络对噪声的鲁棒性。训练过程中排序表损失(ranked list loss, RLL)和分类损失函数共同指导网络训练。RLL能够有效克服传统度量学习损失函数忽略类内特征的问题, 分类损失函数能够弥补度量损失优化下对样本整体分布不敏感的问题。实验表明, 该方法在提高低信噪比雷达信号识别准确率的同时仍具有识别样本库新添加信号类型的能力。

关键词: 雷达信号识别, 深层残差收缩注意力网络, 软阈值化, 注意力机制, 损失函数

Abstract:

Aiming at the low recognition rate of radar signals under the condition of low signal-to-noise ratio, and the classification network has the limitation of identifying the newly added signal types in the sample library, a radar signal recognition method based on the deep residual shrinking attention network is proposed. The one-dimensional radar signal is mapped to a 32-dimensional vector space through the network. The residual connection in the network can effectively strengthen the dissemination ability of features and solve the problem that the network is too deep to be trained; the introduction of the attention mechanism not only constructs the mask branch to act as the feature selector of the main branch, but also helps the network adaptively select the appropriate threshold for soft thresholding, so as to reduce the influence of noise or redundant information in the network and improve the robustness of the network to noise. During the training process, ranked list loss (RLL) and the classification loss function jointly guide the network training. RLL can effectively overcome the problem of traditional metric learning loss function ignoring features within the class, and the classification loss function can make up for the problem of insensitivity to the overall distribution of the sample under metric loss optimization. Experiments show that this method can not only improve the recognition accuracy of low-signal-to-noise ratio radar signals, but also has the ability to identify newly added signal types in the sample library.

Key words: radar signal recognition, deep residuals shrink attention network, soft thresholding, attention mechanism, loss function

中图分类号:

TN974

曹鹏宇, 杨承志, 陈泽盛, 王露, 石礼盟. 基于深度残差收缩注意力网络的雷达信号识别方法[J]. 系统工程与电子技术, 2023, 45(3): 717-725.

Pengyu CAO, Chengzhi YANG, Zesheng CHEN, Lu WANG, Limeng SHI. Radar signal recognition method based on deep residual shrinkage attention network[J]. Systems Engineering and Electronics, 2023, 45(3): 717-725.

图/表 12

图1

图2

图3

图4

表1

表2

图5

图6

图7

图8

图9

图10

参考文献 22

1	WANG Q S , BAI J , HUANG X Y , et al. Analysis of radar emitter signal sorting and recognition model structure[J]. Procedia Computer Science, 2019, 154, 500- 503. doi: 10.1016/j.procs.2019.06.076
2	XU X , WANG W , WANG J H . A three-way incremental-learning algorithm for radar emitter identification[J]. Frontiers of Computer Science, 2016, 10 (4): 673- 688. doi: 10.1007/s11704-015-4457-7
3	ZHANG C H , HAN Y T , ZAHNG P , et al. Research on modern radar emitter modelling technique under complex electromagnetic environment[J]. The Journal of Engineering, 2019, (20): 7134- 7138.
4	HE M, MAO Y, HAN J. A method of extracting radar inpulse characteristics in low SNR[C]//Proc. of the IEEE International Conference on Signal Processing, 2007.
5	SHEN J H , HUANG J C , ZHU Y C , et al. Overview of radar signal fast recognition[J]. Electronic Information Warfare Technology, 2017, 32 (5): 5- 10.
6	LI Y B , GE J , LIN Y , et al. Radar emitter signal recognition based on multi-scale wavelet entropy and feature weighting[J]. Journal of Central South University, 2014, 21 (11): 4254- 4260. doi: 10.1007/s11771-014-2422-5
7	LI L , JI H B . Radar emitter recognition based on cyclostationary signatures and sequential iterative least-square estimation[J]. Expert Systems with Applications, 2011, 38 (3): 2140- 2147. doi: 10.1016/j.eswa.2010.07.155
8	O'SHEA T J , ROY T , 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
9	WANG S H , PHILLIPS P , SUI Y , et al. Classification of Alzheimer's disease based on eight-layer convolutional neural network with leaky rectified linear unit and max pooling[J]. Journal of Medical Systems, 2018, 42 (5): 85. doi: 10.1007/s10916-018-0932-7
10	YUAN X Y , HE P . Adversarial examples: attacks and defenses for deep learning[J]. IEEE Trans.on Neural Networks and Learning Systems, 2019, 30 (9): 2805- 2824. doi: 10.1109/TNNLS.2018.2886017
11	HE K M, ZHANG X Y, REN S Q, et al. Delving deep into rectifiers: surpassing human-level performance on ImageNet classification[C]//Proc. of the IEEE International Conference on Computer Vision, 2015: 1026-1034.
12	刘赢, 田润澜, 王晓峰. 基于深层卷积神经网络和双谱特征的雷达信号识别方法[J]. 系统工程与电子技术, 2019, 41 (9): 1998- 2005.
	LIU Y , TIAN R L , WANG X F . Radar signal recognition method based on deep convolutional neural network and bispectrum feature[J]. Systems Engineering and Electronics, 2019, 41 (9): 1998- 2005.
13	张怡霄, 郭文普, 康凯, 等. 基于聚类和时序相关的重点雷达信号快速识别[J]. 系统工程与电子技术, 2020, 42 (3): 597- 602.
	ZHANG Y X , GUO W P , KANG K , et al. Key radar signal fast recognition method based on clustering and time-series correlation[J]. Systems Engineering and Electronics, 2020, 42 (3): 597- 602.
14	秦鑫, 黄洁, 查雄, 等. 基于扩张残差网络的雷达辐射源信号识别[J]. 电子学报, 2020, 48 (3): 456- 462. doi: 10.3969/j.issn.0372-2112.2020.03.006
	QIN X , HUANG J , ZHA X , et al. Radar emitter signal recognition based on dilated residual network[J]. Acta Electronica Sinica, 2020, 48 (3): 456- 462. doi: 10.3969/j.issn.0372-2112.2020.03.006
15	郭立民, 寇韵涵, 陈涛, 等. 基于栈式稀疏自编码器的低信噪比下低截获概率雷达信号调制类型识别[J]. 电子与信息学报, 2018, 40 (4): 875- 881.
	GUO L M , KOU Y H , CHEN T , et al. Low probability of intercept radar signal recognition based on stacked sparse auto-encoder[J]. Journal of Electronics & Information Technology, 2018, 40 (4): 875- 881.
16	QU Z H , HOU C F , HOU C G , 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. doi: 10.1109/ACCESS.2020.2980363
17	石礼盟, 杨承志, 吴宏超. 基于深层残差网络和三元组损失的雷达信号识别方法[J]. 系统工程与电子技术, 2020, 42 (11): 2506- 2512.
	SHI L M , YANG C Z , WU H C . Radar signal recognition method based on deep residual network and triplet loss[J]. Systems Engineering and Electronics, 2020, 42 (11): 2506- 2512.
18	ZHAO M N , ZHONG S S , FU X Y , et al. Deep residual shrinkage networks for fault diagnosis[J]. IEEE Trans.on Industrial Informatics, 2020, 16 (7): 4681- 4690.
19	WANG F, JIANG M Q, QIAN C, et al. Residual attention network for image classification[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2017.
20	HU J , SHEN L , ALBANIE S , et al. Squeeze-and-excitation networks[J]. IEEE Trans.on Pattern Analysis and Machine Intelligence, 2020, 42 (8): 2011- 2023.
21	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proc. of the IEEE Conference on Computer Vision & Pattern Recognition, 2016.
22	SCHROFF F, KALENICHENKO D, PHILBIN J. FaceNet: a unified embedding for face recognition and clustering[C]//Proc. of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 815-823.

标签	雷达序号	调制方式	信号参数	取值
0	1	CW	载频/MHz	330
1	2	LFM	起始频率/MHz 带宽/MHz	[250, 300] 40
2	3	BPSK	载频/MHz 编码方式	300 13位巴克码
3	4	QPSK	载频/MHz 编码方式	400 格雷码
4	5	P1	载频/MHz 步进频率/MHz	[250, 300] {4, 7}
5	6	P2	载频/MHz 步进频率/MHz	[250, 300] {4, 6}
6	7	P3	载频/MHz 步进频率/MHz	[250, 300] {4, 6, 7}
7	8	P4	载频/MHz 步进频率/MHz	[250, 300] {7, 8}
8	9	P1	载频/MHz 步进频率/MHz	[350, 400] {4, 7}

标签	雷达序号	调制方式	信号参数	取值
-1	10	LFM	起始频率/MHz 带宽/MHz	[200, 250] 35
-1	11	BPSK	载频/MHz 编码方式	250 13位巴克码
-1	12	P1	载频/MHz 步进频率/MHz	[320, 340] {4, 7, 8}
-1	13	P2	载频/MHz 步进频率/MHz	[30, 400] {5, 6}
-1	14	Costas	基准频率调频序列	[125, 150] {3, 4, 6}

[1]	赵庆媛, 赵志强, 叶春茂, 鲁耀兵. 基于自注意力的双波段预警雷达微动融合识别[J]. 系统工程与电子技术, 2023, 45(3): 708-716.
[2]	闫啸家, 梁伟阁, 张钢, 佘博, 田福庆. 基于RCNN-ABiLSTM的机械设备剩余寿命预测方法[J]. 系统工程与电子技术, 2023, 45(3): 931-940.
[3]	韩啸, 陈世文, 陈蒙, 杨锦程. 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.
[4]	赵晓枫, 徐叶斌, 吴飞, 牛家辉, 蔡伟, 张志利. 基于全局感知机制的地面红外目标检测方法[J]. 系统工程与电子技术, 2022, 44(5): 1461-1467.
[5]	徐平亮, 崔亚奇, 熊伟, 熊振宇, 顾祥岐. 生成式中断航迹接续关联方法[J]. 系统工程与电子技术, 2022, 44(5): 1543-1552.
[6]	曹鹏宇, 杨承志, 石礼盟, 吴宏超. 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.
[7]	吴涛, 王伦文, 朱敬成. 基于迁移学习和注意力机制的伪装图像分割[J]. 系统工程与电子技术, 2022, 44(2): 376-384.
[8]	金涛, 王晓峰, 田润澜, 张歆东. 基于改进1DCNN+TCN的雷达辐射源快速识别方法[J]. 系统工程与电子技术, 2022, 44(2): 463-469.
[9]	马钰棠, 孙鹏, 张杰勇, 王鹏, 闫云飞, 赵亮. 样本不平衡下的空中群组意图识别方法[J]. 系统工程与电子技术, 2022, 44(12): 3747-3755.
[10]	唐一强, 杨霄鹏, 朱圣铭. 基于注意力机制的混合CNN-BiLSTM低轨卫星信道预测算法[J]. 系统工程与电子技术, 2022, 44(12): 3863-3870.
[11]	曲凌志, 杨俊安, 刘辉, 黄科举. 嵌入注意力机制的通信辐射源个体识别方法[J]. 系统工程与电子技术, 2022, 44(1): 20-27.
[12]	刘紫燕, 马珊珊, 梁静, 朱明成, 袁磊. 注意力机制CNN的毫米波大规模MIMO系统信道估计算法[J]. 系统工程与电子技术, 2022, 44(1): 307-312.
[13]	崔邦彦, 田润澜, 王东风, 崔钢, 石静苑. 基于注意力机制和改进CLDNN的雷达辐射源识别[J]. 系统工程与电子技术, 2021, 43(5): 1224-1231.
[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.
[15]	方安然, 李旦, 张建秋. 异常值和未知观测噪声鲁棒的卡尔曼滤波器[J]. 系统工程与电子技术, 2021, 43(3): 593-602.