基于多尺度模型空间度量的雷达辐射源行为意图高效识别方法

doi:10.12305/j.issn.1001-506X.2026.08.08

摘要/Abstract

摘要：

准确识别雷达辐射源的行为意图在现代雷达对抗中至关重要。本文提出一种基于多尺度模型空间的雷达行为意图识别方法。通过滑动窗口机制从雷达信号中提取时间序列片段，并进行多尺度分解以获得不同时间尺度的子序列。利用回声状态网络（echo state network，ESN）拟合子序列，捕捉其变化规律，并融合多个ESN模型形成“多尺度ESN模型”，利用该模型表示原始信号，实现从原始数据到模型空间的映射。进一步，引入模型空间的度量算子，在模型空间中寻找最优决策边界，实现精准分类和意图预测。实验显示，所提方法相较于传统方法，显著提高了响应速度和识别精度。

关键词: 模型空间, 辐射源行为意图识别, 回声状态网络

Abstract:

Accurate identification of the behavioral intentions of radar emitters is crucial in modern radar countermeasures. This paper proposes a radar behavioral intention identification method based on a multi-scale model space. By employing a sliding window mechanism, time-series segments are extracted from radar signals and subjected to multi-scale decomposition to obtain subsequences at different time scales. These subsequences are fitted using echo state network (ESN), capturing their patterns of change. Multiple ESN models are integrated to form a “multi-scale ESN model,” representing the original signal and facilitating the mapping from raw data to model space. Additionally, a metric operator in the model space is introduced to identify the optimal decision boundary, enabling precise classification and intention prediction. Experimental results demonstrate that the propsed method significantly enhances response speed and identification accuracy compared to traditional methods.

Key words: model space, radar emitter behavior intention recognition, echo state network (ESN)

中图分类号:

TN 974

段永胜, 薛磊, 徐英, 张峻宁. 基于多尺度模型空间度量的雷达辐射源行为意图高效识别方法[J]. 系统工程与电子技术, 2026, 48(8): 2593-2601.

Yongsheng DUAN, Lei XUE, Ying XU, Junning ZHANG. Efficient behavioral intention identification method of radar emitters based on multi-scale model space metric[J]. Systems Engineering and Electronics, 2026, 48(8): 2593-2601.

图/表 11

图1

图2

表1

表2

表3

表4

表5

表6

图3

图4

表7

参考文献 31

1	ZHAI Q H, LI Y, ZHANG Z L, et al. Few-shot recognition of multi-function radar modes via reﬁned prototypical random walk network[J]. IEEE Trans. on Aerospace and Electronic Systems, 2023, 59 (3): 2376- 2387. doi: 10.1109/TAES.2022.3213792
2	ZHAO Y, WANG X, HUANG Z T. Multi-function radar modeling: a review[J]. IEEE Sensors Journal, 2024, 24 (20): 31658- 31680. doi: 10.1109/JSEN.2024.3436877
3	WANG J , TANG B , WANG C, et al. A novel pulse stream representation and affinity measurement strategy for radar recognition with distribution pattern mining[J]. IEEE Trans. on Instrumentation and Measurement, 2025, 74: 2500514.
4	ZHANG Z L, LI Y, ZHAI Q H, et al. Mode recognition of multifunction radars for few-shot learning based on compound alignments[J]. IEEE Trans. on Aerospace and Electronic Systems, 2022, 58 (6): 5860- 5874. doi: 10.1109/TAES.2022.3180275
5	LIU S K, YAN X P, LI P, et al. Radar emitter recognition based on SIFT position and scale features[J]. IEEE Trans. on Circuits and Systems II: Express Briefs, 2018, 65 (12): 2062- 2066. doi: 10.1109/tcsii.2018.2819666
6	ZHANG M, DIAO M, GAO L P, et al. Neural networks for radar waveform recognition[J]. Symmetry, 2017, 9 (5): 75. doi: 10.3390/sym9050075
7	黄颖坤, 金炜东. 基于SLIDE+SVM的雷达辐射源信号识别[J]. 系统仿真学报, 2017, 29 (9): 1944- 1949.
	HUANG Y K, JIN W D. Radar emitter signal identification based on SLIDE+SVM[J]. Journal of System Simulation, 2017, 29 (9): 1944- 1949.
8	PAN J F, XIONG J W, ZHUO Y H. Working mode recognition of non-specific radar based on ResNet-SVM learning framework[J]. Sensors, 2023, 23 (6): 3123. doi: 10.3390/s23063123
9	XU T, YUAN S, LIU Z M, et al. Radar emitter recognition based on parameter set clustering and classification[J]. Remote Sensing, 2022, 14 (18): 4468. doi: 10.3390/rs14184468
10	DU M Y, ZHONG P, CAI X H, et al. Robust Bayesian attention belief network for radar work mode recognition[J]. Digital Signal Process, 2022, 133, 103874. doi: 10.1016/j.dsp.2022.103874
11	DING L D, WANG S L, WANG F G, et al. Specific emitter identification via convolutional neural networks[J]. IEEE Communications Letters, 2018, 22 (12): 2591- 2594. doi: 10.1109/LCOMM.2018.2871465
12	ZHANG W N, XUE K L, YAO A Q, et al. CTRNet: an automatic modulation recognition based on transformer-CNN neural network[J]. Electronics, 2024, 13 (17): 3408. doi: 10.3390/electronics13173408
13	LIU Z M, YU P S. Classification, denoising and deinterleaving of pulse streams with recurrent neural networks[J]. IEEE Trans. on Aerospace and Electronic Systems, 2019, 55, 1624- 1639. doi: 10.1109/TAES.2018.2874139
14	ZHANG Z W, ZHU M T, LI Y J, et al. Joint recognition and parameter estimation of cognitive radar work modes with LSTM-transformer[J]. Digital Signal Process, 2023, 140, 104081. doi: 10.1016/j.dsp.2023.104081
15	XU X S, BI D P, PAN J F. Method for functional state recognition of multifunction radars based on recurrent neural networks[J]. IET Radar, Sonar & Navigation, 2021, 15(7): 724−732.
16	TANAKA G, YAMANE T, HÉROUX J B, et al. Recent advances in physical reservoir computing: a review[J]. Neural Networks, 2019, 115, 100- 123. doi: 10.1016/j.neunet.2019.03.005
17	LIU S K, ZHOU X R, CHEN H H. From data to D3 model: adaptive subsurface anomaly detection in GPR data[J]. IEEE Trans. on Geoscience and Remote Sensing, 2024, 62, 5105412.
18	VISNEVSKI N. Syntactic modeling of multi-function radars[D]. Hamilton: McMaster University, 2005.
19	贠洁, 孙闽红, 官友廉. 基于雷达字建模的多功能雷达工作模式识别[J]. 杭州电子科技大学学报, 2020, 40 (6): 19- 25.
	YUN J, SUN M H, GUAN Y L. Recognizing multi-function radar working mode based on radar word modeling[J]. Journal of Hangzhou Dianzi University, 2020, 40 (6): 19- 25.
20	FANG Y, BI D P, PAN J F, et al. Muli-fuction radar behavior state detection algorithm based on bayesian critertion[C]//Proc. of the IEEE Advanced Information Technology, Electronic and Automation Control Confernce, 2019.
21	ISHAN B, LONDHE N D. Speaker dependent and independent isolated hindi word recognizer using hidden Markov model （HMM）[J]. International Journal of Computer Applications, 2012, 52 (7): 34- 40. doi: 10.5120/8217-1639
22	CHEN H H, TIŇO P, RODAN A, et al. Learning in the model space for cognitive fault diagnosis[J]. IEEE Trans. on Neural Networks and Learning Systems, 2013, 25 (1): 124- 136.
23	CHEN A, ZHOU X R, FAN Y, et al. Underground diagnosis based on GPR and learning in the model space[J]. IEEE Trans. on Pattern Analysis and Machine Intelligence, 2023, 46 (5): 3832- 3844.
24	SKOLNIK M I. Radar handbook[M]. New York: McGraw-Hill, 2008.
25	马珂, 毕大平, 周圣栾, 等. 基于1DCNN的机载火控雷达空空工作状态识别[J]. 现代雷达, 2023, 45 (3): 17- 23. doi: 10.16592/j.cnki.1004-7859.2023.03.003
	MA K, BI D P, ZHOU S L, et al. Air-to-air operating status recognition of airborne fire control radar based on 1DCNN[J]. Modern Radar, 2023, 45 (3): 17- 23. doi: 10.16592/j.cnki.1004-7859.2023.03.003
26	史新杰. 基于类增量学习的雷达工作模式识别方法研究[D]. 成都: 电子科技大学, 2024.
	SHI X J. Research on radar working mode recognition method based on class incremental learning[D]. Chengdu: University of Electronic Science and Technology of China, 2024.
27	余显祥, 季康龙, 李虎, 等. 基于半监督学习聚类数据标注的多功能雷达工作模式识别[J]. 信息对抗技术, 2023, 2 (6): 29- 46.
	YU X X, JI K L, LI H, et al. Multi-function radar work mode recognition based on clustering data annotation[J]. Information Countermeasure Technology, 2023, 2 (6): 29- 46.
28	王磊, 张志勇, 胥辉旗, 等. 基于VMD分解和多域联合分布的雷达辐射源识别[J]. 系统工程与电子技术, 2023, 45 (8): 2479- 2488.
	WANG L, ZHANG Z Y, XU H Q, et al. Radar emitter identification based on VMD and multi-domain joint distribution[J]. Systems Engineering and Electronics, 2023, 45 (8): 2479- 2488.
29	HUANG Y K, JIN W D, YAN K, et al. Radar emitter signal identification via distance features[J]. Journal of System Simulation, 2021, 33 (12): 2959- 2966.
30	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 and Pattern Recognition, 2016: 770−778.
31	ABDI H, WILLIAMS L J. Principal component analysis[J]. Wiley Interdisciplinary Reviews: Computational Statistics, 2010, 2 (4): 433- 459. doi: 10.1002/wics.101

工作状态	脉冲重复周期/μs	脉冲宽度/μs	占空比/%	单个相干处理间隔内脉冲数	相干处理间隔序列中脉冲重复频率数量
VS（M）	3.3~10.0	1.0~3.0	10.0~30.0	500~2 000	1
VS（E）	3.3~10.0	1.0~3.0	10.0~50.0	500~2 000	1
TWS	50~125.0	1.0~20.0	1.0~25.0	100~256	3~8
TAS（Ⅰ）	50~125.0	1.0~20.0	1.0~50.0	100~256	3~8
TAS（Ⅱ）	50~125.0	0.1~20.0	0.1~50.0	8~150	1~4
STT（M）	50~125.0	0.1~20.0	0.1~25.0	8~150	1~4
STT（E）	50~125.0	0.1~20.0	0.1~50.0	8~150	1~4
DTT	50~125.0	0.1~20.0	0.1~25.0	8~150	1~4

工作模式	工作状态	射频/MHz	脉冲重复周期/μs	脉冲宽度/μs	单个相干处理间隔内脉冲数	调制类型	调制带宽/MHz
VS	搜索	10000 固定	5 固定	1.7	128	线性调频	5
TWS	搜索	10000 固定	67/77/87 组变	1.5	128	线性调频	5
TAS	重搜	10000 固定	11.6/11.8 组变	2.6	64	线性调频	10
	确认	10000 固定	26~26.8 固定	2.6	32	线性调频	10
	跟踪	10000 固定	26~26.4 固定	2.6	32	线性调频	10
RWS	搜索	10000 固定	11.6/11.8 组变	2.6	64	线性调频	10
STT	重搜	10000 固定	52/52.2 组变	1.4	16	线性调频	10
	确认	10000 固定	104~104.8 固定	2.6	8	线性调频	10
	跟踪	10000 固定	104 固定	2.6	8	线性调频	10

算法	识别准确率/%	耗时/s
多尺度ESN + KNN（k=1）	94.64	24.98561
多尺度ESN + KNN（k=3）	95.21	25.13654
多尺度ESN + KNN（k=5）	94.05	25.47382
多尺度ESN + SVM	95.88	27.17529
单尺度ESN + KNN（k=1）	86.77	2.52534
单尺度ESN+ KNN（k=3）	86.70	2.68501
单尺度ESN+ KNN（k=5）	87.19	2.82654
单尺度ESN+ SVM	87.26	3.04550

算法	识别准确率/%	耗时/s
1/3度量 + KNN（k=1）	94.64	24.98561
1/3度量 + KNN（k=3）	95.21	25.13654
1/3度量 + KNN（k=5）	94.05	25.47382
1/3度量 + SVM	95.88	27.17529
欧式距离 + KNN（k=1）	92.88	25.28571
欧式距离 + KNN（k=3）	93.05	25.62349
欧式距离 + KNN（k=5）	92.65	25.94619
欧式距离 + SVM	93.58	27.65453

算法	识别准确率/%	耗时/s
多尺度ESN + KNN（k=3）	95.21	25.13654
多尺度ESN + SVM	95.88	27.17529
DTW + KNN（k=3）	88.35	462.11002
ResNet-18	93.06	282.74875
PCA+SVM	86.20	18.43228