RFFsNet-SEI: a multidimensional balanced-RFFs deep neural network framework for specific emitter identification

doi:10.23919/JSEE.2023.000069

Journal of Systems Engineering and Electronics ›› 2024, Vol. 35 ›› Issue (3): 558-574.doi: 10.23919/JSEE.2023.000069

• HIGH-DIMENSIONAL SIGNAL PROCESSING • Previous Articles

RFFsNet-SEI: a multidimensional balanced-RFFs deep neural network framework for specific emitter identification

Rong FAN¹^,²(), Chengke SI²(), Yi HAN²(), Qun WAN¹^,*()

¹ School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
² Institute of Electronic and Electrical Engineering, Civil Aviation Flight University of China, Guanghan 618307, China

Received:2022-08-10 Accepted:2023-04-22 Online:2024-06-18 Published:2024-06-19
Contact: Qun WAN E-mail:fanrong@cafuc.edu.cn;sichengke@cafuc.edu.cn;Han.Holly@Outlook.com;wanqun@uestc.edu.cn
About author:
FAN Rong was born in 1984. He received his B.S. degree in information engineering from Chengdu University of Technique, Chengdu, China, in 2007, and M.S. and Ph.D. degrees in electronic engineering from University of Electronic Science and Technology of China (UESTC), Chengdu, China, in 2010 and 2014, respectively. In 2014, he joined the 10th Research Institute of China Electronics Technology Group Corporation. In 2018, he joined the Institute of Electronics and Electrical Engineering, Civil Aviation Flight University of China, Guanghan, China. Since 2020, he has been a postdoctor in the UESTC. His research interests include passive detection, direction finding, and specific emitter identification for air traffic control signals. E-mail: fanrong@cafuc.edu.cn

SI Chengke was born in 1988. He received his B.S. and M.S. degrees in computer science from University of Science and Technology of China, Hefei, China, in 2011 and 2014, respectively. In 2014, he joined the 10th Research Institute of China Electronics Technology Group Corporation. In 2019, he joined the Institute of Electronics and Electrical Engineering, Civil Aviation Flight University of China, Guanghan, China. His research interests include deep learning and intelligent sensing of radio signals. E-mail: sichengke@cafuc.edu.cn

HAN Yi was born in 1998. She received her B.S. degree in the Internet of things engineering from Jiangsu University of Technology, Jiangsu, in 2020. She is currently working towards her M.S degree in Institute of Electrical and Electronics Engineering, Civil Aviation Flight University of China, Guanghan, China. Her current research interest is intelligent recognition of UAV radio signals. E-mail: Han.Holly@Outlook.com

WAN Qun was born in 1971. He received his B.S. degree in electronic engineering from Nanjing University, Nanjing, China, in 1993, and M.S. and Ph.D. degrees in electronic engineering from University of Electronic Science and Technology of China (UESTC), Chengdu, China, in 1996 and 2001, respectively. From 2001 to 2003, he was a postdoctor with the Department of Electronic Engineering, Tsinghua University. Since 2004, he has been a professor with the Department of Electronic Engineering, UESTC. He is currently the Director of Joint Research Lab of Array Signal Processing and the Deputy Dean of School of Electronic Engineering. His research interests include direction finding, radio localization, and signal processing based on information criterion. E-mail: wanqun@uestc.edu.cn
Supported by:
The work was supported by the National Natural Science Foundation of China (62061003), Sichuan Science and Technology Program (2021YFG0192), and the Research Foundation of the Civil Aviation Flight University of China (ZJ2020-04;J2020-033).

Abstract

Abstract:

Existing specific emitter identification (SEI) methods based on hand-crafted features have drawbacks of losing feature information and involving multiple processing stages, which reduce the identification accuracy of emitters and complicate the procedures of identification. In this paper, we propose a deep SEI approach via multidimensional feature extraction for radio frequency fingerprints (RFFs), namely, RFFsNet-SEI. Particularly, we extract multidimensional physical RFFs from the received signal by virtue of variational mode decomposition (VMD) and Hilbert transform (HT). The physical RFFs and I-Q data are formed into the balanced-RFFs, which are then used to train RFFsNet-SEI. As introducing model-aided RFFs into neural network, the hybrid-driven scheme including physical features and I-Q data is constructed. It improves physical interpretability of RFFsNet-SEI. Meanwhile, since RFFsNet-SEI identifies individual of emitters from received raw data in end-to-end, it accelerates SEI implementation and simplifies procedures of identification. Moreover, as the temporal features and spectral features of the received signal are both extracted by RFFsNet-SEI, identification accuracy is improved. Finally, we compare RFFsNet-SEI with the counterparts in terms of identification accuracy, computational complexity, and prediction speed. Experimental results illustrate that the proposed method outperforms the counterparts on the basis of simulation dataset and real dataset collected in the anechoic chamber.

Key words: specific emitter identification (SEI), deep learning (DL), radio frequency fingerprint (RFF), multidimensional feature extraction (MFE), variational mode decomposition (VMD)

Rong FAN, Chengke SI, Yi HAN, Qun WAN. RFFsNet-SEI: a multidimensional balanced-RFFs deep neural network framework for specific emitter identification[J]. Journal of Systems Engineering and Electronics, 2024, 35(3): 558-574.

Figures/Tables 26

Fig 1

Fig 2

Fig 3

Fig 4

Fig 5

Fig 6

Fig 7

Fig 8

Fig 9

Fig 10

Table 1

Table 2

Fig 11

Table 3

Table 4

Table 5

Table 6

Table 7

Fig 12

Fig 14

Fig 13

Fig 15

Fig 16

Fig 17

Table 8

Table 9

References 34

1	XU S H. On the identification technique of individual transmitter based on signal prints. Wuhan: Huazhong University of Science and Technology, 2007. (in Chinese)
2	LI D, LIANG Q H, LIU H Q, et al. A novel multidimensional domain deep learning network for SAR ship detection. IEEE Trans. on Geoscience and Remote Sensing, 2022, 60: 5203213.
3	SANKHE K, BELGIOVINE M, ZHOU F, et al No radio left behind: radio fingerprinting through deep learning of physical-layer hardware impairments. IEEE Trans. on Cognitive Communications and Networking, 2020, 6 (1): 165- 178. doi: 10.1109/TCCN.2019.2949308
4	CHEN H, YANG J N, LIU H Communication transmitter individual identification based on deep residual adaptation network. Systems Engineering and Electronics, 2021, 43 (3): 603- 609.
5	YUAN H L, HU A Q Preamble-based detection of Wi-Fi transmitter RF fingerprints. Electronics Letters, 2010, 46 (1): 1165- 1167.
6	SUN L T, WANG X, HUANG Z T, Unintentional modulation microstructure enlargement. Journal of Systems Engineering and Electronics, 2022, 33(3): 522−533.
7	BERTONCINI C, RUDD K, NOUSAIN B, et al Wavelet fingerprinting of radio-frequency identification (RFID) tags. IEEE Trans. on Industrial Electronics, 2005, 59 (12): 4843- 4850.
8	LOPEZ G, GRAJAL J, OSORIO A, et al Digital channelized receiver based on time-frequency analysis for signal interception. IEEE Trans. on Aerospace and Electronic Systems, 2005, 41 (3): 879- 898. doi: 10.1109/TAES.2005.1541437
9	LUNDEN J, KOIVUNEN V. Automatic radar waveform recognition. IEEE Journal of Selected Topics in Signal Processing, 2007, 1(1): 124−136.
10	LIU Z G, ZHANG Q G An approach to recognize the transient disturbances with spectral kurtosis. IEEE Trans. on Instrumentation and Measurement, 2014, 63 (1): 46- 55. doi: 10.1109/TIM.2013.2277513
11	YUAN Y J, HUANG Z T, WU H, et al Specific emitter identification based on hilbert-huang transform-based time-frequency-energy distribution features. IET Communication, 2014, 8 (13): 2404- 2412. doi: 10.1049/iet-com.2013.0865
12	HE B, WANG F G Cooperative specific emitter identification via multiple distorted receivers. IEEE Trans. on Information Forensics and Security, 2016, 15, 3791- 3806.
13	SATIJA U, TRIVEDI N, BISWAL G, et al Specific emitter identification based on variational mode decomposition and spectral features in single hop and relaying scenarios. IEEE Trans. on Information Forensics and Security, 2019, 14 (3): 581- 591. doi: 10.1109/TIFS.2018.2855665
14	SUN L T, WANG X, YANG A F, et al Radio frequency fingerprint extraction based on multi-dimension approximate entropy. IEEE Signal Processing Letters, 2020, 27, 471- 475. doi: 10.1109/LSP.2020.2978333
15	GOODFELLOW I, BENGIO Y, COURVILLE A. Deep learning: adaptive computation and machine learning series. Cambridge: MIT Press, 2016.
16	WANG Y, GUI G, GACANIN H, et al An efficient specific emitter identification method based on complex-valued neural networks and network compression. IEEE Journal on Selected Areas in Communications, 2021, 39 (8): 2305- 2317. doi: 10.1109/JSAC.2021.3087243
17	QU L Z, LIU H, HUANG K J Specific emitter identification based on multi-domain feature fusion and integrated learning. Symmetry, 2021, 13 (8): 1481. doi: 10.3390/sym13081481
18	DING L, WANG S L, WANG F G, et al Specific emitter identification via convolutional neural networks. IEEE Communications Letters, 2018, 22 (12): 2591- 2594. doi: 10.1109/LCOMM.2018.2871465
19	XIAO Y, WEI X Z Specific emitter identification of radar based on one dimensional convolution neural network. Journal of Physics: Conference Series, 2020, 1550, 032114. doi: 10.1088/1742-6596/1550/3/032114
20	WENG L T R, PENG J L, HE Y, et al Specific emitter identification of ADS-B signal based on deep residual network. Aero Weaponry, 2021, 28 (4): 24- 29.
21	XIAO Z L, YAN Z Y Radar emitter identification based on novel time-frequency spectrum and convolutional neural network. IEEE Communications Letters, 2021, 25 (8): 2634- 2638. doi: 10.1109/LCOMM.2021.3084043
22	LIU Z M Multi-feature fusion for specific emitter identification via deep ensemble learning. Digital Signal Processing, 2021, 110, 102939. doi: 10.1016/j.dsp.2020.102939
23	ZHAO Y R, WANG X, LIN Z Y, et al Multi-classifier fusion for open-set specific emitter identification. Remote Sensing, 2022, 14, 2226. doi: 10.3390/rs14092226
24	ZHUO F, HUANG Y L, CHEN J. Specific emitter identification based on linear polynomial fitting of the energy envelope. Proc. of the IEEE 6th International Conference on Electronics Information and Emergency Communication, 2016: 278−281.
25	ZHANG J, WANG F, DOBRE O A Specific emitter identification via Hilbert-Huang transform in single-hop and relaying scenarios. IEEE Trans. on Information Forensics Security, 2016, 11 (6): 1192- 1205. doi: 10.1109/TIFS.2016.2520908
26	WONG L J, HEADLEY S, ANDREWS R , et al. Clustering learned CNN features from raw I/Q data for emitter identification. Proc. of the IEEE Military Communications Conference, 2018: 26−33.
27	IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift. Proc. of the International Conference on Machine Learning, 2015: 448−456.
28	FAN R, ZHU X Y, TANG W B, et al MFFNet: multi-path features fusion network for source enumeration. IEEE Communications Letters, 2022, 26 (3): 572- 576. doi: 10.1109/LCOMM.2021.3139676
29	GONG J L, XU X D, LEI Y K Unsupervised specific emitter identification method using radio-frequency fingerprint embedded InfoGAN. IEEE Trans. on Information Forensics and Security, 2020, 15, 2898- 2913. doi: 10.1109/TIFS.2020.2978620
30	FAN R, SI C K, GUO H S, et al Direction finding for coherent sources with deep hybrid neural networks. International Journal of Electronics, 2022, 109 (5): 811- 833. doi: 10.1080/00207217.2021.1941293
31	KENINGTON P B. High linearity RF amplifier design. Boston: Artech House, 2000.
32	HU J, SHEN L, ALBANIE S, et al Squeeze-and-excitation networks. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2020, 42 (8): 2011- 2023. doi: 10.1109/TPAMI.2019.2913372
33	ZHANG C B, JIANG P T, HOU Q B, et al Delving deep into label smoothing. IEEE Trans. on Image Processing, 2021, 30, 5984- 5996. doi: 10.1109/TIP.2021.3089942
34	SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research, 2014, 15 (1): 1929- 1958.

Name	Component	Output size	CBL	Conv	#Res
ResBlock1	ResUnit(top)	64×1×16	3×1,16/1	3×1,16/1	1×1,16/1
	ResUnit(mid.)	64×1×16	3×1,16/1	3×1,16/1	1×1,16/1
	ResUnit(bot.)	64×1×16	3×1,16/1	3×1,16/1	1×1,16/1
CBFs	Component	Output size	CBL(top)	CBL(mid.)	CBL(bot.)
	CBF(left)	64×1×1	1×1,16/1	3×1,32/1	1×1,1/1
	CBF(mid.)	64×1×3	1×1,16/1	3×1,32/1	1×1,3/1
	CBF(right)	64×1×4	1×1,16/1	3×1,32/1	1×1,4/1
	CBL	64×1×8	1×1,8/1	−	−
ResBlock2	Component	Output size	CBL	Conv	#Res
	ResUnit(top)	32×1×16	3×1,16/2	3×1,16/1	1×1,16/2
	ResUnit(mid.)	16×1×16	3×1,16/2	3×1,16/1	1×1,16/2
	ResUnit(bot.)	8×1×16	3×1,16/2	3×1,16/1	1×1,16/2
CBF*s	Component	Output size	CBL(top)	CBL(mid.)	CBL(bot.)
	CBF*(left)	1×1×2	1×1,16/2	3×1,32/2	1×1,2/2
	CBF*(mid.)	1×1×3	1×1,16/2	3×1,32/2	1×1,3/2
	CBF*(right)	1×1×3	1×1,16/2	3×1,32/2	1×1,3/2
	CBL	1×1×8	1×1,8/1	−	−
CBF	Component	Output size	CBL(top)	CBL(mid.)	CBL(bot.)
CBF	CBF	1×1×8	1×1,16/1	3×1,32/1	1×1,8/1
	FC	10×1	−	−	−

Emitter	V=2			U=3
Emitter	$ {h_0} $	$ {h_1} $	$ {h_2} $	$ {\chi _1} $	$ {\chi _3} $	$ {\chi _5} $
E1	1.00	−0.36	−0.36	1.00	−0.36	−0.36
E2	1.00	−0.27	−0.27	1.00	−0.27	−0.27
E3	1.00	−0.18	−0.18	1.00	−0.18	−0.18
E4	1.00	−0.09	−0.09	1.00	−0.09	−0.09
E5	1.00	0.00	0.00	1.00	0.00	0.00
E6	1.00	0.09	0.09	1.00	0.09	0.09
E7	1.00	0.18	0.18	1.00	0.18	0.18
E8	1.00	0.27	0.27	1.00	0.27	0.27
E9	1.00	0.36	0.36	1.00	0.36	0.36
E10	1.00	0.45	0.45	1.00	0.45	0.45

Type	Power output/dBm	Waves type	Pulse repetition interval/μs	Duty ratio	RF output/MHz
E4438C	0	Pulse wide modulation	50	26/50	1575

Frequency range	Power supply voltage	Maximum power output/dBm	Gain	System impedance/Ω
0.1 MHz−2 GHz	6−12 VDC	+10	25 dB @ 1500 MHz	50

Frequency range	Tuning frequency/MHz	Bandwidth/MHz	Attenuation/dB	Work mode	IF value/MHz
30 MHz−3 GHz	1575	4	0	normal	70

RFFsNet-SEI: a multidimensional balanced-RFFs deep neural network framework for specific emitter identification

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

Share this article

Figures/Tables 26

References 34

Related Articles 3

Recommended Articles

Metrics

Comments

Sampling rate	Quantization bit	DDC value	Sampling duration/s	I-Q rates/MHz
250 MSPS	16	2	5	125

Step	Note	MFLOPs	Para.	FPS	Simulation-data A_cc/%	Real-data A_cc/%
S1	RFFsNet-SEI-Basic	601.5	12976	15366	49.1	21.0
S2	+ ResBlcoks1	1756.5	18384	10380	58.3	39.2
S3	+ResBlcoks2	1993.4	23872	7135	77.2	78.1
S4	+auxiliary CBL blocks	2627.2	26730	5645	79.7	82.3
S5	n_C:16→8	692.0	7514	9851	76.0	78.1
S6	n_C:16→32	10266.8	100874	2687	79.6	82.3

Number of IMFs	RFFSNET-SEI	VMD-KNN
2	(6 002, 78.7%)	(99, 56.7%)
3	(5 875, 79.1%)	(91, 59.2%)
4	(5 645, 79.7%)	(84, 61.5%)
5	(5 284, 79.4%)	(79, 60.9%)

Criterion	RFFS-SVM [11]	VMD-KNN [13]	MAPEN [14]	MDFFIL [17]	BISPECTRAL-CNN [18]	CNN-1D [19]	DRN [20]	MFFDEL-SEI [22]	RFFSNET-SEI (Proposed)
FPS	171	84	144	413	372	35643	8969	1073	5 645
A_cc/%	64.0	61.5	13.3	73.7	45.3	79.6	72.3	76.6	79.7
MFLOPs	−	−	−	14318	77536	54	1684	2160	2627
Parameters	−	−	−	719066	1246314	28112	105884	258856	26730

[1]	Siting LYU, Xiaohui LI, Tao FAN, Jiawen LIU, Mingli SHI. Deep learning for fast channel estimation in millimeter-wave MIMO systems [J]. Journal of Systems Engineering and Electronics, 2022, 33(6): 1088-1095.
[2]	Liting SUN, Xiang WANG, Zhitao HUANG. Unintentional modulation microstructure enlargement [J]. Journal of Systems Engineering and Electronics, 2022, 33(3): 522-533.
[3]	Chuan LIN, Qing CHANG, Xianxu LI. Uplink NOMA signal transmission with convolutional neural networks approach [J]. Journal of Systems Engineering and Electronics, 2020, 31(5): 890-898.