Radar emitter signal recognition method based on improved collaborative semi-supervised learning

doi:10.23919/JSEE.2023.000126

Abstract

Abstract:

Rare labeled data are difficult to recognize by using conventional methods in the process of radar emitter recognition. To solve this problem, an optimized cooperative semi-supervised learning radar emitter recognition method based on a small amount of labeled data is developed. First, a small amount of labeled data are randomly sampled by using the bootstrap method, loss functions for three common deep learning networks are improved, the uniform distribution and cross-entropy function are combined to reduce the overconfidence of softmax classification. Subsequently, the dataset obtained after sampling is adopted to train three improved networks so as to build the initial model. In addition, the unlabeled data are preliminarily screened through dynamic time warping (DTW) and then input into the initial model trained previously for judgment. If the judgment results of two or more networks are consistent, the unlabeled data are labeled and put into the labeled data set. Lastly, the three network models are input into the labeled dataset for training, and the final model is built. As revealed by the simulation results, the semi-supervised learning method adopted in this paper is capable of exploiting a small amount of labeled data and basically achieving the accuracy of labeled data recognition.

Key words: emitter signal identification, time series, bootstrap, semi supervised learning, cross entropy function, homogenization, dynamic time warping (DTW)

Tao JIN, Xindong ZHANG. Radar emitter signal recognition method based on improved collaborative semi-supervised learning[J]. Journal of Systems Engineering and Electronics, 2023, 34(5): 1182-1190.

Figures/Tables 15

Fig 1

Fig 2

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

Table 8

Table 9

Fig 3

Fig 4

Fig 5

Fig 6

References 30

1	WANG S Q, BAI J, HUANG X Y, et al Analysis of radar emitter signal sorting and recognition model structure. Procedia Computer Science, 2019, 154, 500- 503. doi: 10.1016/j.procs.2019.06.076
2	MAN P, DING C B, REN W J, et al A nonlinear fingerprint-level radar simulation modeling method for specific emitter identification. Electronics, 2021, 10 (9): 1030. doi: 10.3390/electronics10091030
3	CAO R, CAO J W, MEI J P, et al Radar emitter identification with bispectrum and hierarchical extreme learning machine. Multimedia Tools and Applications, 2019, 78 (20): 28953- 28970. doi: 10.1007/s11042-018-6134-y
4	MENG F J, TANG H, WANG Y Z, et al Radar radiation source signal recognition based on texture characteristics of time-frequency images. Journal of Missile and Guidance, 2017, 37 (3): 152- 156.
5	CAO X H, WANG L X, SHU X Y Radar emitter signal recognition based on wavelet moment invariants. Computer Engineering and Applications, 2020, 56 (19): 269- 272.
6	LI Y B, GE J, LIN Y, et, al Radar emitter signal recognition based on multi-scale wavelet entropy and feature weighting. Journal of Central South University, 2014, 21 (11): 4254- 4260. doi: 10.1007/s11771-014-2422-5
7	KONG S H, KIM M, HOANG L M Automatic LPI radar waveform recognition using CNN. IEEE Access, 2018, 6, 4207- 4219. doi: 10.1109/ACCESS.2017.2788942
8	VAN ENGELEN J E, HOOS H H A survey on semi-supervised learning. Machine Learning, 2020, 109 (2): 373- 440. doi: 10.1007/s10994-019-05855-6
9	OUALI Y, HUDELOT C, TAMI M. An overview of deep semi-supervised learning. http://arxiv.org/abs/2006.05278.
10	FENG Y T, WANG G L, LIU Z P, et al An unknown radar emitter identification method based on semi-supervised and transfer learning. Algorithms, 2019, 12 (12): 271. doi: 10.3390/a12120271
11	MOSTAJABI M, WANG C M, RANJAN D, et al. High-resolution radar dataset for semi-supervised learning of dynamic objects. Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020: 100−101.
12	YING F, XING W Radar signal recognition based on modified semi-supervised SVM algorithm. Proc. of the 2nd Advanced Information Technology, Electronic and Automation Control Conference, 2017, 2336- 2340.
13	MULLER M. Information retrieval for music and motion. Berlin: Springer, 2007: 69−84.
14	PERMANASARI Y, HARAHAP E H, ALI E P Speech recognition using dynamic time warping (DTW). Journal of Physics: Conference Series, 2019, 1366 (1): 012091.
15	BERNDT D J, CLIFFORD J. Using dynamic time warping to find patterns in time series. Proc. of the 3rd International Conference on Knowledge Discovery and Data Mining, 1994: 359−370.
16	SENIN P. Dynamic time warping algorithm review. http://api.semanticscholar.org/CorpusID:16629907.
17	BARLOW H B Unsupervised learning. Neural Computation, 1989, 1 (3): 295- 311. doi: 10.1162/neco.1989.1.3.295
18	GHAHRAMANI Z Unsupervised learning. Proc. of the Summer School on Machine Learning, 2003, 72- 112.
19	SHAHAM T R, DEKEL T, MICHAELI T. Singan: learning a generative model from a single natural image. Proc. of the IEEE/CVF International Conference on Computer Vision, 2019: 4570−4580.
20	LU X J Automatic segmentation and structural characterization of low density fibreboards. Image Analysis & Stereology, 2013, 32 (1): 13- 25.
21	CHONG Y W, DING Y, YAN Q, et al Graph-based semi-supervised learning: a review. Neurocomputing, 2020, 408, 216- 230. doi: 10.1016/j.neucom.2019.12.130
22	DUNLOP M M, SLEPCEV D, STUART A M, et al Large data and zero noise limits of graph-based semi-supervised learning algorithms. Applied and Computational Harmonic Analysis, 2020, 49 (2): 655- 697. doi: 10.1016/j.acha.2019.03.005
23	SUBRAMANYA A, TALUKDAR P P. Synthesis lectures on artificial intelligence and machine learning. Cham: Springer, 2014.
24	ZHUANG C, MA Q. Dual graph convolutional networks for graph-based semi-supervised classification. Proc. of the World Wide Web Conference, 2018: 499−508.
25	ZHAI X L, ZHOU Z H, TIN C Semi-supervised learning for ECG classification without patient-specific labeled data. Expert Systems with Applications, 2020, 158, 113411. doi: 10.1016/j.eswa.2020.113411
26	SUN Y G, XU J Q, WU H, et al. Deep learning based semi-supervised control for vertical security of maglev vehicle with guaranteed bounded airgap. IEEE Trans. on Intelligent Transportation Systems, 2021, 22(7): 4431−4442.
27	LU G, WANG Y B, YANG H Y, et al One-dimensional convolutional neural networks for acoustic waste sorting. Journal of Cleaner Production, 2020, 271, 122393. doi: 10.1016/j.jclepro.2020.122393
28	HAN A G, BYRA M, HEBA E, et al Noninvasive diagnosis of nonalcoholic fatty liver disease and quantification of liver fat with radiofrequency ultrasound data using one-dimensional convolutional neural networks. Radiology, 2020, 295 (2): 342- 350. doi: 10.1148/radiol.2020191160
29	FARHA Y A, GALL J. MS-TCN: multi-stage temporal convolutional network for action segmentation. Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019: 3575−3584.
30	KOK C, JAHMUNAH V, OH S L, et al Automated prediction of sepsis using temporal convolutional network. Computers in Biology and Medicine, 2020, 127, 103957. doi: 10.1016/j.compbiomed.2020.103957

Attribute	Content
Algorithm name	Improved cooperative semi-supervised learning classification algorithm
Classification object	Eight kinds of radar radiation source sequence data
Algorithm objectives	Use a small amount of labeled data and a large amount of unlabeled data to achieve the effect of supervised recognition
Major innovations	Improving the activation function to improve the recognition accuracy, while reducing the network time consumption without reducing the classification accuracy by DTW technique
Classification networks	CNN, TCN, CTA networks
Algorithm flow	The similarity between labeled and unlabeled data is calculated in groups for filtering, and then the initially trained classification network model is used to judge each unlabeled data

Signal	Main parameter	Value
BPSK	Barker code digits	{7,11,13}
Multi-phase code	Number of code bits	{36,64}
FMCW	Sampling frequency/kHz	{15,17}
	Modulation period/ms	{50,25,35}
	Modulation bandwidth/kHz	{0.25,0.35,0.5}
Costas	Frequency sequences/kHz	{[4 7 1 6 5 2 3], [2 6 3 8 7 5 1]}

Network	Loss 1	Accuracy 1	Loss 2	Accuracy 2
CTA	1.7665	0.7316	0.6915	0.8480
CNN	1.2586	0.7420	1.2987	0.8274
TCN	1.4941	0.7846	1.8574	0.8501

Network	Loss 1	Accuracy 1	Loss 2	Accuracy 2
CTA	1.0657	0.8046	0.6615	0.9339
CNN	1.0477	0.8261	0.6866	0.9271
TCN	0.9493	0.8578	0.6616	0.9344

Number of labeled data	CTA	CNN	TCN
64	0.6532	0.7520	0.7523
64	0.8356	0.8323	0.8361
128	0.8046	0.8261	0.8578
128	0.9339	0.9271	0.9344
256	0.8527	0.8846	0.9030
256	0.9397	0.9345	0.9403