基于注意力机制的堆叠LSTM网络雷达HRRP序列目标识别方法

doi:10.12305/j.issn.1001-506X.2021.10.09

摘要/Abstract

摘要：

传统的雷达高分辨距离像(high resolution range profile, HRRP)序列识别方法依赖于人工特征提取, 并且现有的深度学习方法存在梯度消失问题, 导致收敛速度慢, 识别精度低。针对上述问题, 提出一种基于注意力机制的堆叠长短时记忆(attention-based stacked long short-term memory, Attention-SLSTM)网络模型, 该模型通过堆叠多个长短时记忆(long short-term memory, LSTM)网络层, 实现了HRRP序列更深层次抽象特征的提取; 通过替换模型的激活函数, 减缓了堆叠LSTM(stacked LSTM, SLSTM)模型梯度消失问题; 引入注意力机制计算特征序列的分配权重并用于分类识别步骤, 增强了隐藏层特征的非线性表达能力。模型在雷达目标识别标准数据集MSTAR上多种不同目的的实验结果表明, 所提方法具有更快的收敛速度和更好的识别性能, 与多种现有方法对比具有更高的识别率, 证明了所提方法的正确性和有效性。

关键词: 高分辨距离像序列, 注意力机制, 长短时记忆网络, 雷达目标识别

Abstract:

The traditional radar high resolution range profile (HRRP) sequence recognition method relies on artificial feature extraction, and the existing deep learning method has the problem of gradient vanishing, which leads to the slow convergence speed and low recognition accuracy of the existing recognition methods. To solve these problems, an attention-based stacked long short-term memory (Attention-SLSTM) network model is proposed, which realizes the extraction of deeper abstract features of HRRP sequence by stacking multiple long short-term memory (LSTM) network layers.By replacing the activation function of the model, it slows down the gradient vanishing problem of stacked LSTM.The attention mechanism is introduced to calculate the distribution weight of feature sequence and use it in the classification and recognition step, which enhances the nonlinear expression ability of hidden layer features. Experimental results on the radar target recongnition standard data set MSTAR for different purposes show that the proposed method has faster convergence speed and better recognition performance, and has higher recognition rate compared with other existing methods, which proves the correctness and effectiveness of the proposed method.

Key words: high-resolution range profile sequence (HRRPs), attention mechanism, long short-term memory (LSTM) network, radar automatic target recognition (RATR)

中图分类号:

TN95

张一凡, 张双辉, 刘永祥, 荆锋. 基于注意力机制的堆叠LSTM网络雷达HRRP序列目标识别方法[J]. 系统工程与电子技术, 2021, 43(10): 2775-2781.

Yifan ZHANG, Shuanghui ZHANG, Yongxiang LIU, Feng JING. Radar HRRP sequence target recognition method of attention mechanism based stacked LSTM network[J]. Systems Engineering and Electronics, 2021, 43(10): 2775-2781.

图/表 11

图1

图2

图3

图4

图5

图6

表1

图7

表2

表3

图8

参考文献 31

1	DU L , WANG P H , LIU H W , et al. Bayesian spatiotemporal multitask learning for radar HRRP target recognition[J]. IEEE Trans.on Signal Processing, 2011, 59 (7): 3182- 3196. doi: 10.1109/TSP.2011.2141664
2	刘天坤, 熊新平, 赵育善. 网络化防空导弹体系生存能力的评估与权衡[J]. 系统工程与电技术, 2007, 29 (2): 226- 229.
	LIU T K , XIONG X P , ZHAO Y S . Research on survivability evaluation and tradeoff of networking air defense missile systems[J]. Systems Engineering and Electronics, 2007, 29 (2): 226- 229.
3	CAPLAN P , SHIMIZU H , KOBAYSHI S , et al. Cataloging internet resources[J]. The Public Access Computer Systems Review, 1993, 4 (2): 61- 66.
4	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): 577- 602.
5	徐彬, 陈渤, 刘家麒, 等. 采用双向LSTM模型的雷达HRRP目标识别[J]. 西安电子科技大学学报, 2019, 46 (4): 29- 34.
	XU B , CHEN B , LIU J Q , et al. Radar HRRP target recognition using bidirectional LSTM model[J]. Journal of Xidian University, 2019, 46 (4): 29- 34.
6	DU L , LIU H W , WANG P H , et al. Radar HRRP target recognition by the bidirectional LSTM model[J]. IEEE Trans.on Signal Processing, 2012, 60 (7): 3546- 3559. doi: 10.1109/TSP.2012.2191965
7	WANG C Y , HUANG P P , LI X F , et al. Radar HRRP target recognition based on AEPSO-SVM algorithm[J]. Systems Engineering and Electronics, 2019, 41 (9): 1984- 1989.
8	LAN D, WANG P H, LIU H W, et al. Radar HRRP target recognition based on dynamic multi-task hidden Markov model[C]// Proc. of the IEEE International Conference on National Radar, 2011: 253-255.
9	王彩云, 黄盼盼, 李晓飞, 等. 基于AEPSO-SVM算法的雷达HRRP目标识别[J]. 系统工程与电子技术, 2019, 41 (9): 1984- 1989.
	WANG C Y , HUANG P P , LI X F , et al. Radar HRRP target recognition based on AEPSO-SVM algorithm[J]. Systems Engineering and Electronics, 2019, 41 (9): 1984- 1989.
10	DEWI C , CHEN R C , HENDRY H , et al. Experiment improvement of restricted Boltzmann machine methods for image classification[J]. Vietnam Journal of Computer Science, 2021, 8 (2): 134- 141.
11	HSU K C , LI S T . Clustering spatial-temporal precipitation data using wavelet transform and self-organizing map neural network[J]. Advances in Water Resources, 2010, 33 (2): 190- 200. doi: 10.1016/j.advwatres.2009.11.005
12	GUO Y, WANG S, GAO C, et al. Wishart RBM based DBN for polarimetric synthetic radar data classification[C]//Proc. of the IEEE International Geoscience and Remote Sensing Symposium, 2015.
13	刘凯, 张斌, 黄青华. 基于TCNN-BiLSTM网络的调制识别算法[J]. 系统工程与电子技术, 2020, 42 (8): 1841- 1849.
	LIU K , ZHANG B , HUANG Q H . Modulation recognition algorithm based on TCNN-BiLSTM[J]. Systems Engineering and Electronics, 2020, 42 (8): 1841- 1849.
14	GUO C , WANG H , JIANT T , et al. Radar target recognition based on feature pyramid fusion lightweight CNN[J]. IEEE Access, 2019, 7, 51140- 51149. doi: 10.1109/ACCESS.2019.2909348
15	ZHANG Y F , XIAO F , QIAN F C , et al. VGM-RNN: HRRP sequence extrapolation and recognition based on a novel optimized RNN[J]. IEEE Access, 2020, 8, 70071- 70081. doi: 10.1109/ACCESS.2020.2986027
16	LIU J, CHEN B, CHEN W, et al. Radar HRRP target recognition with target aware two-dimensional recurrent neural net work[C]//Proc. of the IEEE International Conference on Signal Processing, Communications and Computing (ICSPCC), 2020.
17	LI X, HE Y, YANG Y, et al. LSTM based human activity classification on radar range Profile[C]//Proc. of the IEEE International Conference on Computational Electromagnetics, 2019.
18	WAN J W , CHEN B , LIU Y Q , et al. Recognizing the HRRP by combining CNN and BiRNN with attention mechanism[J]. IEEE Access, 2020, 8, 20828- 20837. doi: 10.1109/ACCESS.2020.2969450
19	WANG Y, HUANG M, ZHU X, et al. Attention-based LSTM for aspect-level sentiment classification[C]//Proc. of the Conference on Empirical Methods in Natural Language Processing, 2016.
20	YAN S, SMITH J S, LU W, et al. CHAM: action recognition using convolutional hierarchical attention model[C]//Proc. of the IEEE International Conference on Image Processing (ICIP), 2017.
21	WANG A H , CHEN J P , ZHANG Y Z , et al. Comparison of three widely used multi-GNSS real-time single-frequency precise point positioning models using the international GNSS service real-time service[J]. IET Radar Sonar Navigation, 2020, 14 (1): 1- 14. doi: 10.1049/iet-rsn.2020.0009
22	DU C , ZHANG L , SUN X P , et al. Enhanced multi-channel feature synthesis for hand gesture recognition based on CNN with a channel and spatial attention mechanism[J]. IEEE Access, 2020, 8, 14610- 14620.
23	CUI Z Y , WANG X Y , LIU N Y , et al. Ship detection in large-scale SAR images via spatial shuffle-group enhance attention[J]. IEEE Trans.on Geoscience and Remote Sensing, 2020, 59 (1): 379- 391.
24	ALBRECHT T W, GUSTAFSON S C. Hidden Markov models for classifying SAR target images[C]//Proc. of the Algorithms for Synthetic Aperture Radar Imagery XI. International Society for Optics and Photonics, 2004.
25	LIANG Z, HUANG J, LIU J P, et al. Enterprise master patient index entity recognition by long short-term memory network in electronic health systems[C]//Proc. of the British HCI 2018, 2018.
26	SHRESTHA A , LI H , KERNEC J L , et al. Continuous human activity classification from FMCW radar with Bi-LSTM networks[J]. IEEE Sensors Journal, 2020, 22 (20): 13607- 13619.
27	L I X, HE Y, YANG Y M, et al. LSTM based human activity classification on radar range profile[C]//Proc. of the IEEE Internation al Conference on Computational Electromagnetics, 2019.
28	TIAN B , WANG G , XU Z , et al. Communication delay compensation for string stability of CACC system using LSTM prediction[J]. Vehicular Communications, 2021, 100333.
29	LI X , LIU Z , HUANG Z . Attention-based radar PRI modulation recognition with recurrent neural networks[J]. IEEE Access, 2020, 8, 57426- 57436. doi: 10.1109/ACCESS.2020.2982654
30	REN H , YU X , ZOU L , et al. Extended convolutional capsule network with application on SAR automatic target recognition[J]. Signal Processing, 2021, 183 (1): 108021.
31	SUN Z , LIU M , LIU P , et al. SAR Image classification using fully connected conditional random fields combined with deep learning and superpixel boundary constraint[J]. Remote Sensing, 2021, doi: 10.3390/rs13020271

类别	训练集(17°)	测试集(15°)
BMP2(SN-9566)	2 330	1 960
BTR70(SN-C71)	2 330	1 960
T72(SN-132)	2 320	1 960
BTR60	2 560	1 950
2S1	2 990	2 740
BRDM2	2 980	2 740
D7	2 990	2 740
T62	2 990	2 730
ZIL131	2 990	2 740
ZSU23/4	2 990	2 740
合计	27 470	32 030

层数	模型	识别率
1	SVM	0.820 0
1	RNN	0.867 0
1	LSTM	0.882 4
1	SLSTM-sigmoid	0.890 9
1	SLSTM-Elu	0.905 4
1	Attention-RNN	0.910 5
1	Attention-SLSTM	0.948 0

层数	SLSTM- sigmoid	SLSTM- Elu	Attention- RNN	Attention- SLSTM
1	0.890 9	0.905 4	0.910 5	0.948 0
2	0.918 9	0.925 5	0.929 4	0.952 7
3	0.928 2	0.936 0	0.942 2	0.969 0
4	0.902 2	0.943 4	0.954 6	0.972 4
5	0.887 7	0.950 2	0.960 0	0.978 6
6	0.882 4	0.954 3	0.958 3	0.980 0
7	0.880 0	0.942 2	0.952 5	0.980 3
8	0.872 5	0.903 0	0.952 1	0.978 8
9	0.868 0	0.892 0	0.951 5	0.978 5
10	0.865 4	0.880 5	0.952 0	0.979 0

[1]	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.
[2]	王彩云, 吴钇达, 王佳宁, 马璐, 赵焕玥. 基于改进的CNN和数据增强的SAR目标识别[J]. 系统工程与电子技术, 2022, 44(8): 2483-2487.
[3]	孙晶明, 虞盛康, 孙俊. 基于元学习的雷达小样本目标识别方法及改进[J]. 系统工程与电子技术, 2022, 44(6): 1839-1845.
[4]	吉瑞萍, 张程祎, 梁彦, 王跃东. 基于LSTM的弹道导弹主动段轨迹预报[J]. 系统工程与电子技术, 2022, 44(6): 1968-1976.
[5]	侯召国, 王华伟, 周良, 付强. 基于改进深度残差网络的旋转机械故障诊断[J]. 系统工程与电子技术, 2022, 44(6): 2051-2059.
[6]	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.
[7]	孙晶明, 虞盛康, 孙俊. 基于深度学习的HRRP识别姿态敏感性分析[J]. 系统工程与电子技术, 2022, 44(3): 802-807.
[8]	孙世岩, 张钢, 梁伟阁, 佘博, 田福庆. 基于时间序列数据扩增和BLSTM的滚动轴承剩余寿命预测方法[J]. 系统工程与电子技术, 2022, 44(3): 1060-1068.
[9]	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.
[10]	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.
[11]	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.
[12]	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.
[13]	刘旗, 张新禹, 刘永祥. 基于门控多尺度匹配网络的小样本SAR目标识别[J]. 系统工程与电子技术, 2022, 44(11): 3346-3356.
[14]	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.
[15]	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.