基于集成学习与特征降维的小样本调制识别方法

doi:10.12305/j.issn.1001-506X.2021.04.28

摘要/Abstract

摘要：

针对有标签样本较少条件下的通信信号调制识别问题, 提出一种基于集成学习与特征降维的小样本调制方式分类模型。首先，通过集成人工特征与深度学习自动提取特征构成特征集合。然后，设计特征选择算法对特征合集进行优选生成高效特征子集。最后, 利用可快速收敛的高性能分类器对信号进行区分, 实现在少量有标签样本和大量无标签样本条件下的调制方式分类。仿真结果表明, 通过对8种数字信号进行调制识别, 在信噪比为20 dB时, 所提算法可将信号最高识别率提升至96%, 同时该算法设计简单, 具有较大应用价值。

关键词: 调制识别, 小样本, 集成学习, 特征选择

Abstract:

Aiming at the problem of communication signal modulation recognition with few labeled samples, a few-shot modulation classification model based on ensemble learning and feature dimension reduction is proposed. Firstly, the feature set is formed by integrating handcrafted features and deep learning features. And then, the feature selection algorithm is designed to optimize the feature set to generate an optimal feature subset. Finally, the signals are distinguished by a fast convergent high-performance classifier, which can realize the modulation classification under the condition of a small number of labeled samples and a large number of unlabeled samples. The simulation results show that when the signal to noise ratio is 20 dB, the proposed algorithm can improve the signal recognition rate to 96% through modulation recognition of eight kinds of digital signals. At the same time, the algorithm is simple and has great application value.

Key words: modulation recognition, few-shot, ensemble learning, feature selection

中图分类号:

TN975

史蕴豪, 许华, 郑万泽, 刘英辉. 基于集成学习与特征降维的小样本调制识别方法[J]. 系统工程与电子技术, 2021, 43(4): 1099-1109.

Yunhao SHI, Hua XU, Wanze ZHENG, Yinghui LIU. Few-shot modulation recognition method based on ensemble learning and feature dimension reduction[J]. Systems Engineering and Electronics, 2021, 43(4): 1099-1109.

图/表 15

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

表1

图11

图12

图13

表2

参考文献 27

1	KHARBECH S , DAYOUB I , SIMON E , et al. On classifiers for blind feature-based automatic modulation classification over MIMO Channels[J]. IET Communications, 2016, 10 (7): 790- 795. doi: 10.1049/iet-com.2015.1124
2	ZHE X , YONG G . Method to reduce the signal-to-noise ratio required for modulation recognition based on logarithmic properties[J]. IET Communications, 2018, 12 (11): 1360- 1366. doi: 10.1049/iet-com.2018.0030
3	GHASEMI S, GANGAL A. An effective algorithm for automatic modulation recognition[C]//Proc. of the IEEE 22nd Conference on Signal Processing and Communications Applications, 2014: 903-906.
4	SHERME A E . A novel method for auto-matic modulation recognition[J]. Applied Soft Computing, 2012, 12 (1): 453- 461. doi: 10.1016/j.asoc.2011.08.025
5	WANG H, GUO L L. A new method of automatic modulation recognition based on dimension reduction[C]//Proc. of the IEEE Forum on Cooperative Positioning and Service, 2017: 316-320.
6	ZHANG Z , LI Y B , JIN S S , et al. Modulation signal recognition based on information entropy and ensemble learning[J]. Entropy, 2018, 20 (3): 198- 198. doi: 10.3390/e20030198
7	WANG H , GUO L L , DOU Z , et al. A new method of cognitive signal recognition based on hybrid information entropy and D-S evidence theory[J]. Mobile Networks and Applications, 2018, 23 (4): 677- 685. doi: 10.1007/s11036-018-1000-8
8	O'SHEA T J, CORGAN J, CLANCY T C, et al. Convolutional radio modulation recognition networks[C]//Proc. of the International Conference on Engineering Applications of Neural Networks, 2016: 213-226.
9	O'SHEA T J , ROY T , CLANCY T C , et al. 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
10	JEONG S, LEE U, KIM S C. Spectrogram-based automatic modu-lation recognition using convolution neural network[C]//Proc. of the IEEE 10th International Conference on Ubiquitous and Future Networks, 2018: 843-845.
11	FAN M , PENG C , LENAN W , et al. Automatic modulation classification: a deep learning enabled approach[J]. IEEE Trans.on Vehicular Technology, 2018, 67 (11): 10760- 10772. doi: 10.1109/TVT.2018.2868698
12	ZHANG Z , WANG C , GAN C , et al. Automatic modulation classification using convolutional neural network with features fusion of SPVWD and BJD[J]. IEEE Trans.on Signal and Information Processing over Networks, 2019, 16 (4): 1- 15.
13	李跃, 郭兴吉, 赵欣. 基于高阶累积量的调制方式识别研究[J]. 西南科技大学学报, 2018, 33 (3): 64- 68. doi: 10.3969/j.issn.1671-8755.2018.03.013
	LI Y , GUO X J , ZHAO X . Study on modulation recognition based on higher-order cumulants[J]. Journal of Southwest University of Science and Technology, 2018, 33 (3): 64- 68. doi: 10.3969/j.issn.1671-8755.2018.03.013
14	谭晓衡, 褚国星, 张雪静, 等. 基于高阶累积量和小波变换的调制识别算法[J]. 系统工程与电子技术, 2018, 40 (1): 171- 177.
	TAN X H , ZHU G X , ZHANG X J , et al. Modulation recognition algorithm based on highorder-cumulants and wavelet transform[J]. Systems Engineering and Electronics, 2018, 40 (1): 171- 177.
15	李晨, 杨俊安, 刘辉. 基于信息熵和GA-ELM的调制识别算法[J]. 系统工程与电子技术, 2020, 42 (1): 223- 229.
	LI C , YANG J A , LIU H . Modulation recognition algorithm based on information entropy and GA-ELM[J]. Systems Engineering and Electronics, 2020, 42 (1): 223- 229.
16	ALI A , FAN Y Y . Automatic modulation classification using deep learning based on sparse autoencoders with non-negativity constraints[J]. IEEE Signal Processing Letter, 2017, 24 (11): 1626- 1630. doi: 10.1109/LSP.2017.2752459
17	YA T, LIN Y, WANG H. Modulation recognition of digital signal based on deep auto-ancoder network[C]//Proc. of the IEEE International Conference on Software Quality, Reliability and Security Companion, 2017: 256-260.
18	O'SHEA T J, CORGAN J, CLANCY T C. Unsupervised representation learning of structured radio communication signals[C]//Proc. of the Learning for Intelligent Machines, 2016.
19	ASLAM M W , ZHU Z , NANDI A K . Automatic modulation classification using combination of genetic programming and KNN[J]. IEEE Trans.on Wireless Communication, 2012, 11 (8): 2742- 2750.
20	ANG J C, HARON H, NUZLY H, et al. Semi-supervised SVM-based feature selection for cancer classification using microarray gene expression data[C]//Proc. of the International Conference on Industrial, 2015: 468-477.
21	HAN Y H , YANG Y , YAN Y , et al. Semi-supervised feature selection via spline regression for video semantic recognition[J]. IEEE Trans.on Neural Networks Learning System, 2015, 26 (6): 252- 264.
22	BENABDESLEM K , HINDAWI M . Efficient semi-supervised feature selection: constraint, relevance, and redundancy[J]. IEEE Trans. on Knowledge and Data Engineering, 2014, 26 (5): 1131- 1143. doi: 10.1109/TKDE.2013.86
23	SHEIKHPOUR R , SARRAM M A , GHARAGHANI S , et al. A survey on semi-supervised feature selection methods[J]. Pattern Recognition, 2016, 31 (3): 35- 45.
24	YU L , LIU H . Efficient feature selection via analysis of relevance and redundancy[J]. Journal of Machine Learning Research, 2004, 5 (12): 1205- 1224.
25	WANG Y T , WANG J D , LIAO H . An efficient semi-supervised representatives feature selection algorithm based on information theory[J]. Pattern Recognition, 2017, 61 (5): 11- 52.
26	PENG H , LONG F , DING C . Feature selection based on mutual information: criteria of max-dependency, max-relevance, and min-redundancy[J]. IEEE Trans.on Pattern Analysis and Machine Intelligence, 2005, 27 (4): 1226- 1238.
27	ZHANG S C , LI X L . Learning K for KNN classification[J]. ACM Trans.on Intelligent System and Technology, 2017, 8 (8): 1- 19.

特征选择算法	所选特征数	特征子集	时间/s
mRMR	26	f₂、f₃、f₅、f₆、f₇、f₈、f₉、f₁₀、f₁₁、f₁₂、f₁₄、f₁₅、f₁₆、f₁₉、f₂₁、f₂₂、f₂₇、f₂₈、f₃₀、f₃₁、f₃₂、f₃₃、f₃₄、f₃₅、f₃₆、f₃₇	0.7
FCBF	19	f₃、f₆、f₇、f₈、f₉、f₁₀、f₁₂、f₁₄、f₁₅、f₁₆、f₁₉、f₂₁、f₂₇、f₂₈、f₃₀、f₃₂、f₃₄、f₃₆、f₃₇	1.3
SRFS	7	f₃、f₉、f₁₂、f₁₆、f₂₆、f₂₇、f₃₆	37.4
本文算法	13	f₃、f₇、f₈、f₉、f₁₀、f₁₂、f₁₆、f₁₉、f₂₁、f₂₇、f₃₂、f₃₆、f₃₇	9.8

分类器	最高识别准确率/%	训练时间/s
BP网络	96.0	7.04
XGBOOST	94.1	6.26
KNN	91.2	8.31

[1]	李波, 周家豪, 刘民岷, 朱品朝. 基于改进NSGA3的焊接缺陷评估特征选择[J]. 系统工程与电子技术, 2022, 44(7): 2211-2218.
[2]	康颖, 赵治华, 吴灏, 李亚星, 孟进. 基于Deep SVDD的通信信号异常检测方法[J]. 系统工程与电子技术, 2022, 44(7): 2319-2328.
[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]	秦博伟, 蒋磊, 许华, 齐子森. 基于残差生成对抗网络的调制识别算法[J]. 系统工程与电子技术, 2022, 44(6): 2019-2026.
[5]	邵凯, 朱苗苗, 王光宇. 基于生成对抗与卷积神经网络的调制识别方法[J]. 系统工程与电子技术, 2022, 44(3): 1036-1043.
[6]	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.
[7]	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.
[8]	马玉凤, 向南, 豆亚杰, 姜江, 杨克巍, 谭跃进. 军事系统工程中的知识图谱应用及研究[J]. 系统工程与电子技术, 2022, 44(1): 146-153.
[9]	马骏, 杨镜宇, 邹立岩. 基于Stacking集成元模型的作战体系能力图谱生成方法[J]. 系统工程与电子技术, 2022, 44(1): 154-163.
[10]	曹浪财, 林晓昌, 苏思行. 基于矩阵分解和自适应图的无监督特征选择[J]. 系统工程与电子技术, 2021, 43(8): 2197-2208.
[11]	李睿峰, 许爱强, 孙伟超, 王树友. 基于元学习的航空电子设备特征选择算法推荐方法[J]. 系统工程与电子技术, 2021, 43(7): 2011-2020.
[12]	丁嘉辉, 汤建龙, 于正洋. 轻量化的增量式集成学习算法设计[J]. 系统工程与电子技术, 2021, 43(4): 861-867.
[13]	晏媛, 孙俊, 孙晶明, 于俊朋. 雷达小样本目标识别方法及应用分析[J]. 系统工程与电子技术, 2021, 43(3): 684-692.
[14]	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.
[15]	曾强, 黄政, 魏曙寰. 融合专家先验知识和单调性约束的贝叶斯网络参数学习方法[J]. 系统工程与电子技术, 2020, 42(3): 646-652.