用于雷达信号分选的连通k近邻聚类算法

doi:10.12305/j.issn.1001-506X.2023.08.20

摘要/Abstract

摘要：

为了能够在密集且复杂多变的信号环境中进行实时有效的信号分选, 需要一种具有较低复杂度且能够根据信号环境自动调整参数的聚类方法。在模糊聚类算法的基础上结合k近邻搜索, 将λ邻域范围搜索变为λ邻域内k近邻搜索, 提出了连通k近邻聚类(connected k-nearest neighbor clustering, CkNNC)算法。相比模糊聚类算法, 所提算法时间复杂度降低而空间复杂度稍有增加。为使得该算法能够根据信号环境自动进行参数调整, 提出了基于k距离图的阈值参数确定方法。所提算法具有时间复杂度低与阈值参数自动确定的特点, 仿真结果表明所提算法与使用Calinski-Harabasz指标确定最佳阈值的低复杂度模糊聚类算法相比，分选效果差距不大、性能相近, 而时间复杂度大幅下降。

关键词: 电子对抗, 信号分选, 聚类, k近邻, k距离图

Abstract:

To perform real-time and effective signal sorting in a dense, complex, changeable signal environment, a clustering method with lower complexity and capable of automatically adjusting parameters according to the signal environment is required. Based on the fuzzy clustering algorithm, combined with k-nearest neighbor search, the λ-neighborhood search is changed to the k-nearest neighbor search in the λ-neighborhood, and a connected k-nearest neighbor clustering (CkNNC) algorithm is proposed. Compared with the fuzzy clustering algorithm, the proposed algorithm's time complexity is reduced while the proposed algorithm's space complexity is slightly increased. In order to enable the algorithm to automatically adjust parameters according to the signal environment, a threshold parameter determination method based on the k-distance graph is proposed. The proposed algorithm has the characteristics of low time complexity and automatic determination of threshold parameters. Simulation results show that the proposed algorithm is not far from the low-complexity fuzzy clustering algorithm that uses the Calinski-Harabasz index to determine the best threshold, and the performance is similar, and the time complexity is greatly reduced.

Key words: electronic countermeasures, signal sorting, clustering, k-nearest neighbor, k-distance graph

中图分类号:

TN957.51

司伟建, 张悦, 邓志安. 用于雷达信号分选的连通k近邻聚类算法[J]. 系统工程与电子技术, 2023, 45(8): 2463-2470.

Weijian SI, Yue ZHANG, Zhian DENG. Connected k-nearest neighbor clustering algorithm for radar signal sorting[J]. Systems Engineering and Electronics, 2023, 45(8): 2463-2470.

图/表 16

图1

图2

图3

图4

图5

表1

图6

图7

图8

图9

表2

表3

表4

图10

表5

表6

参考文献 26

1	汪飞, 刘建锋. 编队雷达的脉冲重复间隔低分选设计[J]. 信息技术, 2019, (3): 14-18, 23
	WANG F , LIU J F . LPS design of pulse repetition interval for formation radars[J]. Information Technology, 2019, (3): 14-18, 23
2	张保群. 一种抗SDIF分选的脉冲重复间隔参差设计方法[J]. 兵器装备工程学报, 2016, 37 (9): 87-91, 114
	ZHANG B Q . A design method of PRI stagger countering the SDIF sorting algorithm[J]. Journal of Ordnance Equipment Engineering, 2016, 37 (9): 87-91, 114
3	冒燕, 何明浩, 朱元清. 基于PRI参数的雷达信号抗分选分析[J]. 微计算机信息, 2006, (4): 260-261, 236 doi: 10.3969/j.issn.1008-0570.2006.04.095
	MAO Y , HE M H , ZHU Y Q . Analysis of radar signal counter-deinterleaving based on PRI parameter[J]. Microcomputer Information, 2006, (4): 260-261, 236 doi: 10.3969/j.issn.1008-0570.2006.04.095
4	吴惟诚, 潘继飞, 杨丽. 雷达信号分选技术研究综述[J]. 飞航导弹, 2016, (12): 71- 76.
	WU W C , PAN J F , YANG L . Summary of research on radar signal sorting technology[J]. Aerodynamic Missile Journal, 2016, (12): 71- 76.
5	TAN P N , STEINBACH M , KARPATNE A , et al. Introduction to data mining[M]. Global ed Harlow: Pearson Education, 2019.
6	彭刚, 袁晓, 刘闻. 雷达辐射源信号聚类分选算法综述[J]. 雷达科学与技术, 2019, 17 (5): 485- 492.
	PENG G , YUAN X , LIU W . A survey of clustering and sorting algorithms for radar source signals[J]. Radar Science and Technology, 2019, 17 (5): 485- 492.
7	孙鑫, 侯慧群, 杨承志. 基于改进K-均值算法的未知雷达信号分选[J]. 现代电子技术, 2010, 33 (17): 91-93, 96
	SUN X , HOU H Q , YANG C Z . Unknown radar signals deinterleaving based on improved K-means algorithm[J]. Mo-dern Electronics Technique, 2010, 33 (17): 91-93, 96
8	井塬塬. 基于C均值聚类算法的雷达信号分选方法[J]. 电子科技, 2011, 24 (4): 4- 7.
	JING Y Y . Radar signal sorting method based on C-means clustering[J]. Electronic Science and Technology, 2011, 24 (4): 4- 7.
9	陈利虎, 张尔扬, 沈荣骏. 基于优化初始聚类中心K-Means算法的跳频信号分选[J]. 国防科技大学学报, 2009, 31 (2): 70- 75.
	CHEN L H , ZHANG E Y , SHEN R J . The sorting of frequency hopping signals based on K-Means algorithm with optimal initial clustering centers[J]. Journal of National University of Defense Technology, 2009, 31 (2): 70- 75.
10	SU S Q, FU X J, ZHAO C X, et al. Unsupervised k-means combined with SOFM structure adaptive radar signal sorting algorithm[C]//Proc. of the IEEE International Conference on Signal, Information and Data Processing, 2019.
11	翁永祥, 赵万磊, 郑志娟. 基于PCA联合K-Means聚类的雷达信号分选算法[J]. 舰船电子对抗, 2020, 43 (2): 38-42, 60
	WENG Y X , ZHAO W L , ZHENG Z J . Radar signal sorting algorithm based on PCA combined K-means clustering[J]. Shipboard Electronic Countermeasure, 2020, 43 (2): 38-42, 60
12	丁增斗. 基于SFCM聚类算法的雷达信号分选方法[J]. 舰船电子对抗, 2019, 42 (5): 79- 82.
	DING Z D . Radar signal sorting method based on SFCM clus-tering algorithm[J]. Shipboard Electronic Countermeasure, 2019, 42 (5): 79- 82.
13	CAO S, WANG S C, ZHANG Y. Density-based fuzzy C-means multi-center re-clustering radar signal sorting algorithm[C]//Proc. of the IEEE 17th International Conference on Machine Learning and Applications, 2018: 891-896.
14	ESTER M, KRIEGEL H P, SANDER J, et al. A density-based algorithm for discovering clusters in large spatial databases with noise[C]//Proc. of the 2nd International Conference on Knowledge Discovery and Data Mining, 1996: 226-231.
15	李宗林, 罗可. DBSCAN算法中参数的自适应确定[J]. 计算机工程与应用, 2016, 52 (3): 70-73, 80
	LI Z L , LUO K . Research on adaptive parameters determination in DBSCAN algorithm[J]. Computer Engineering and Applications, 2016, 52 (3): 70-73, 80
16	王光, 林国宇. 改进的自适应参数DBSCAN聚类算法[J]. 计算机工程与应用, 2020, 56 (14): 45- 51.
	WANG G , LIN G Y . Improved adaptive parameter DBSCAN clustering algorithm[J]. Computer Engineering and Applications, 2020, 56 (14): 45- 51.
17	李文杰, 闫世强, 蒋莹, 等. 自适应确定DBSCAN算法参数的算法研究[J]. 计算机工程与应用, 2019, 55 (5): 1-7, 148
	LI W J , YAN S Q , JIANG Y , et al. Research on method of self-adaptive determination of DBSCAN algorithm parameters[J]. Computer Engineering and Applications, 2019, 55 (5): 1-7, 148
18	万佳, 胡大裟, 蒋玉明. 多密度自适应确定DBSCAN算法参数的算法研究[J]. 计算机工程与应用, 2022, 58 (2): 78- 85.
	WAN J , HU D S , JIANG Y M . Research on method of multi-density self-adaptive determination of DBSCAN algorithm parameters[J]. Computer Engineering and Applications, 2022, 58 (2): 78- 85.
19	王星, 陈相, 周一鹏, 等. 一种基于改进DBSCAN的雷达信号分选算法[J]. 空军工程大学学报(自然科学版), 2021, 22 (3): 47- 54.
	WANG X , CHEN X , ZHOU Y P , et al. A radar signal sorting algorithm based on improved DBSCAN algorithm[J]. Journal of Air Force Engineering University (Natural Science Edition), 2021, 22 (3): 47- 54.
20	CHEN X, LIU D, WANG X, et al. Improved DBSCAN radar signal sorting algorithm based on rough set[C]//Proc. of the IEEE 2nd International Conference on Big Data and Informatization Education, 2021: 398-401.
21	张兴华. 模糊聚类分析的新算法[J]. 数学的实践与认识, 2005, (3): 138- 141.
	ZHANG X H . A New algorithm for fuzzy clustering[J]. Mathematics in Practice and Theory, 2005, (3): 138- 141.
22	贺宏洲, 景占荣, 徐振华. 雷达信号的模糊聚类分选方法[J]. 航空计算技术, 2008, 38 (5): 21- 24.
	HE H Z , JING Z R , XU Z H . Fuzzy cluster sorting method of radar signals[J]. Aeronautical Computing Technique, 2008, 38 (5): 21- 24.
23	刘旭波, 司锡才. 基于改进的模糊聚类的雷达信号分选[J]. 弹箭与制导学报, 2009, 29 (5): 278- 282.
	LIU X B , SI X C . Sorting of radar-signals based on modified fuzzy clustering[J]. Journal of Projectiles, Rockets, Missiles and Guidance, 2009, 29 (5): 278- 282.
24	尹亮, 潘继飞, 姜秋喜. 基于模糊聚类的雷达信号分选[J]. 火力与指挥控制, 2014, 39 (2): 52-54, 57
	YIN L , PAN J F , JIANG Q X . A study on sorting of radar-signals based on fuzzy clustering[J]. Fire Control & Command Control, 2014, 39 (2): 52-54, 57
25	张悦, 司伟建. 基于并查集的低复杂度模糊聚类信号分选算法[J]. 电波科学学报, 2021, 36 (5): 797- 806.
	ZHANG Y , SI W J . Low complexity fuzzy clustering signal sorting algorithm based on disjoint set[J]. Chinese Journal of Radio Science, 2021, 36 (5): 797- 806.
26	ARYA S , MOUNT D M , NETANYAHU N S , et al. An optimal algorithm for approximate nearest neighbor searching fixed dimensions[J]. Journal of the ACM, 1998, 45 (6): 891- 923.

编号	PRI类型	方位角/(°)	俯仰角/(°)	载频/MHz	脉宽/μs	PRI/μs
1	常规	3	-6	5 000	30	300
2	参差	-2	-10	10 000	18	70/105/150
3	常规	-15	-10	3 000	20	200
4	参差	-17	-5	12 000	5	50/60/80
5	10%抖动	-13	-7	2 000	40	600
6	常规	-18	-8	4 000/4 050/4 100	30	200
7	参差	20	-5	15 000	10	100/130/160/210
8	常规	21	-2	10 000	30	700
9	常规	23	-7	16 000/16 050/16 100	40	600
10	10%抖动	17	-4	17 500	20	600

编号	信号数量	分选所得信号数量	正确分选信号数量	准确率
1	334	338	334	0.988 2
2	923	925	923	0.997 8
3	500	503	500	0.994 0
4	1 578	1 581	1 578	0.998 1
5	167	167	167	1.000 0
6	500	505	500	0.990 1
7	667	671	667	0.994 0
8	143	146	143	0.979 5
9	167	167	167	1.000 0
10	167	168	167	0.994 0

雷达	次数
雷达	1	2	3	4	5
雷达1	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达2	0.975 7	0.984 5	0.997 8	0.985 6	0.987 8
雷达3	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达4	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达5	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达6	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达7	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达8	0.991 7	1.000 0	0.942 9	1.000 0	0.961 5
雷达9	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达10	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
平均	0.996 7	0.998 4	0.994 1	0.998 6	0.994 9
平均	0.996 6

雷达	次数
雷达	1	2	3	4	5
雷达1	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达2	0.975 7	0.984 5	1.000 0	0.985 6	0.986 7
雷达3	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达4	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达5	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达6	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达7	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达8	0.991 7	0.992 2	1.000 0	1.000 0	0.953 8
雷达9	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
雷达10	1.000 0	1.000 0	1.000 0	1.000 0	1.000 0
平均	0.996 7	0.997 7	1.000 0	0.998 6	0.994 1
平均	0.997 4

编号	PRI类型	方位角/(°)	俯仰角/(°)	载频/ MHz	脉宽/ μs	PRI/μs
1	常规	10	5	3 000	7	100
2	10%抖动	-10	10	6 000	40	700
3	5%抖动	8	5	12 000	5	500
4	参差	-5	10	8 000	15	90/130/160
5	常规	0	-3	8 800/9 000/9 200	40	2 000