基于A-DBSCAN的欠定盲源分离算法

doi:10.3969/j.issn.1001-506X.2020.12.02

摘要/Abstract

摘要：

为提升欠定盲源分离问题中混合矩阵的估计精度,在噪声环境下基于密度的空间聚类(density-based spatial clustering of applications with noise, DBSCAN)算法的基础上,提出一种自适应确定输入参数的DBSCAN算法(adaptive DBSCAN, A-DBSCAN)用于混合矩阵估计。针对DBSCAN算法邻域半径(Eps)及邻域点数(MinPts)依赖人为设定的问题,首先利用曲线拟合方法得出Eps,然后通过分析聚类输出类别数与噪声点数关系确定MinPts,并将其与混合矩阵估计模型相结合,最后通过最短路径算法实现源信号恢复。实验结果表明,提出的算法在估计混合矩阵和恢复源信号时,相关性能与对比算法相较均有明显提升。

关键词: 欠定盲源分离, 密度聚类, 曲线拟合, 邻域半径, 邻域点数

Abstract:

In order to improve the estimation accuracy of the mixing matrix in the underdetermined blind source separation, an adaptive density-based spatial clustering of applications with noise (A-DBSCAN) algorithm is proposed, which is based on the DBSCAN algorithm. To solve the problem that the neighborhood radius (Eps) and the number of neighborhood points (MinPts) of the DBSCAN algorithm are determined, First, the curve fitting method is used to obtain Eps, then MinPts is determined by analyzing the relationship between the number of cluster output categories and the number of noise points. The proposed algorithm is combined with the mixing matrix estimation model, and finally the source signal recovery is achieved by the shortest path algorithm. Experimental results show that compared with the comparison algorithm, the proposed algorithm has significantly improved the performance of mixing matrix estimation and source signal recovery.

Key words: underdetermined blind source separation, density clustering, curve fitting, neighborhood radius, neighborhood point

中图分类号:

TN911.7

季策, 穆文欢, 耿蓉. 基于A-DBSCAN的欠定盲源分离算法[J]. 系统工程与电子技术, 2020, 42(12): 2676-2683.

Ce JI, Wenhuan MU, Rong GENG. Underdetermined blind source separation algorithm based on A-DBSCAN[J]. Systems Engineering and Electronics, 2020, 42(12): 2676-2683.

图/表 14

图1

图2

图3

图4

图5

图6

图7

图8

表1

图9

图10

图11

表2

表3

参考文献 25

1	BERTRAND A , MOONEN M . Distributed canonical correlation analysis in wireless sensor networks with application to distributed blind source separation[J]. IEEE Trans.on Signal Proces-sing, 2015, 63 (18): 4800- 4813. doi: 10.1109/TSP.2015.2443729
2	YE F , CHEN J , GAO L P , et al. A mixing matrix estimation algorithm for the time-delayed mixing model of the underdetermined blind source separation problem[J]. Circuits Systems and Signal Processing, 2019, 38 (4): 1889- 1906. doi: 10.1007/s00034-018-0930-5
3	ZHANG L J, YE Z, ZHANG Y, et al. Underdetermined mixing matrix estimation based on single source detection[C]//Proc.of the China International SAR Symposium, 2018.
4	KARIM D , SALAH H , ADNEN C , et al. Hybridization DE with K-means for speaker clustering in speaker diarization of broadcasts news[J]. International Journal of Speech Technology, 2019, 22 (4): 893- 909. doi: 10.1007/s10772-019-09633-6
5	GALAN S F . Comparative evaluation of region query strategies for DBSCAN clustering[J]. Information Sciences, 2019, 502, 76- 90. doi: 10.1016/j.ins.2019.06.036
6	LIN C Y , CHANG C C , LIN C C . A new density-based scheme for clustering based on genetic algorithm[J]. Fundamental Information, 2005, 68 (4): 315- 331.
7	杨静, 高嘉伟, 梁吉业. 基于数据场的改进DBSCAN聚类算法[J]. 计算机科学与探索, 2012, 6 (10): 903- 911.
	YANG J , GAO J W , LIANG J Y . Improved DBSCAN clustering algorithm based on data field[J]. Computer Science and Exploration, 2012, 6 (10): 903- 911.
8	冯振华, 钱雪忠, 赵娜娜. Greedy DBSCAN:一种针对多密度聚类的DBSCAN改进算法[J]. 计算机应用研究, 2016, 33 (9): 2693- 2696, 2700.
	FENG Z H , QIAN X Z , ZHAO N N . Greedy DBSCAN: an improved DBSCAN algorithm for multi-density clustering[J]. Computer Application Research, 2016, 33 (9): 2693- 2696, 2700.
9	TOHTAYONG M, KHAN S, YAACOB M, et al. Harmonics elimination based on curve fitting method for modular multilevel converters[C]//Proc.of the IEEE Region 10 Conference, 2017: 1890-1895.
10	LI Y N, PENG W B, ZHU X Y, et al. Based on short-time fourier transform impulse frequency response analysis in the application of the transformer winding deformation[C]//Proc.of the IEEE International Conference on Mechatronics and Automation, 2016: 371-375.
11	任喜顺, 沈越泓, 高猛, 等. 基于时频分析的混合矩阵估计方法[J]. 信号处理, 2012, 28 (4): 545- 553.
	REN X S , SHEN Y H , GAO M , et al. Mixing matrix estimation method based on time frequency analysis[J]. Signal Processing, 2012, 28 (4): 545- 553.
12	GUO Q , RUAN G Q , NAN P L . Underdetermined mixing matrix estimation algorithm based on single source points[J]. Circuits Systems and Signal Processing, 2017, 36 (11): 4453- 4467. doi: 10.1007/s00034-017-0522-9
13	谭北海, 赵敏, 谢胜利. 基于无监督学习的源数估计及盲分离算法[J]. 系统工程与电子技术, 2009, 31 (8): 1790- 1794.
	TAN B H , ZHAO M , XIE S L . Source number estimation and blind separation algorithm based on unsupervised learning[J]. Systems Engineering and Electronics, 2009, 31 (8): 1790- 1794.
14	BRYANT A C , CIOS K J . RNN-DBSCAN: a density-based clustering algorithm using reverse nearest neighbor density estimates[J]. IEEE Trans.on Knowledge and Data Engineering, 2018, 30 (6): 1109- 1121. doi: 10.1109/TKDE.2017.2787640
15	SUN J D , LI Y X , WEN J T , et al. Novel mixing matrix estimation approach in underdetermined blind source separation[J]. Neurocomputing, 2016, 173, 623- 632. doi: 10.1016/j.neucom.2015.08.008
16	董耀刚, 颉丽, 姚正学, 等. 基于傅里叶函数拟合的滑坡变形监测分析[J]. 中国水土保持, 2019, 11, 72- 75.
	DONG Y G , JIE L , YAO Z X , et al. Monitoring analysis of landslide deformation based on Fourier function fitting[J]. Soil and Water Conservation in China, 2019, 11, 72- 75.
17	WANG R, HUANG R, GAO Y, et al. Sampling method for GMM curve fitting in multi object tracking[C]//Proc.of the IEEE International Conference on Imaging Systems and Techniques, 2017.
18	张永涛, 贾延明. 最小二乘法中代数多项式曲线拟合的分析及实现[J]. 计算机与数字工程, 2017, 45 (4): 637- 639, 654.
	ZHANG Y T , JIA Y M . Analysis and realization of algebraic polynomial curve fitting in least square method[J]. Computer and Digital Engineering, 2017, 45 (4): 637- 639, 654.
19	周治平, 王杰锋, 朱书伟, 等. 一种改进的自适应快速AF-DBSCAN聚类算法[J]. 智能系统学报, 2016, 11 (1): 93- 98.
	ZHOU Z P , WANG J F , ZHU S W , et al. An improved adaptive fast AF-DBSCAN clustering algorithm[J]. Journal of Intelligent Systems, 2016, 11 (1): 93- 98.
20	何选森, 何帆. 密度空间与密度峰值聚类的欠定混合矩阵估计[J]. 数据采集与处理, 2019, 34 (5): 819- 830.
	HE X S , HE F . Underdetermined mixed matrix estimation of density space and density peak clustering[J]. Data Acquisition and Processing, 2019, 34 (5): 819- 830.
21	LYU S, LIU C C, ZHANG C, et al. Underdetermined blind speech signal separation method based on the improved shortest path method[C]//Proc.of the International Conference on Swarm Intelligence, 2015: 397-404.
22	TAHERI N, KACHENOURA A, ANSARIASL K, et al. Feasibility of blind source separation methods for the denoising of dense-array EEG[C]//Proc.of the IEEE 37th Annual International Conference of the Engineering in Medicine & Biology Society, 2015: 4773-4776.
23	ZHEN L , PENG D , YI Z , et al. Underdetermined blind source separation using sparse coding[J]. IEEE Trans.on Neural Networks and Learning Systems, 2017, 28 (12): 3102- 3108. doi: 10.1109/TNNLS.2016.2610960
24	WANG W, SU D L, ZHAO Z H. Mixing matrix estimation in blind source separation based on CFSFDP[C]//Proc.of the International Applied Computational Electromagnetics Society Symposium, 2019.
25	LEI X J , DING Y L , WU F X . Detecting protein complexes from DIPNs by density based clustering with pigeon-inspired optimization algorithm[J]. Science China Information Sciences, 2016, 59 (7): 070103. doi: 10.1007/s11432-016-5578-9

算法	Eps	MinPts	聚类数	F值
DBSCAN	0.005	4	6	0.727 6
	0.010	4	4	0.738 4
	0.015	4	3	0.899 2
A-DBSCAN	0.007	8	3	0.953 7

实验2	混合矩阵	NMSE/dB
算法1	$\left[ \begin{array}{l}\begin{array}{{20}{c}}{0.441\;5}&{ - 0.725\;1}&{0.924\;0}\end{array}\\\begin{array}{{20}{c}}{0.897\;2}&{\;\;0.688\;6}&{\;0.382\;4}\end{array}\end{array} \right]$	-31.827 6
算法2	$\left[ \begin{array}{l}\begin{array}{{20}{c}}{0.407\;7}&{ - 0.712\;0}&{0.926\;8}\end{array}\\\begin{array}{{20}{c}}{0.913\;1}&{\;\;0.702\;2}&{\;0.375\;5}\end{array}\end{array} \right]$	-46.732 6
算法3	$\left[ {\begin{array}{{20}{l}}{\begin{array}{{20}{c}}{0.409\;0}&{ - 0.708\;2}&{0.926\;9}\end{array}}\\{\begin{array}{*{20}{c}}{0.912\;5}&{\;\;0.706\;0}&{\;0.375\;2}\end{array}}\end{array}} \right]$	-55.314 3

实验2	算法1	算法2	算法3
源信号1	0.906 2	0.911 4	0.911 7
源信号2	0.956 8	0.959 5	0.959 8
源信号3	0.968 9	0.971 2	0.971 4

[1]	曹鹏宇, 杨承志, 石礼盟, 吴宏超. 基于PSO-DBSCAN和SCGAN的未知雷达信号处理方法[J]. 系统工程与电子技术, 2022, 44(4): 1158-1165.
[2]	赵孟晨, 姚秀娟, 王静, 董苏惠. 基于Stacked-TCN的空间混叠信号单通道盲源分离方法[J]. 系统工程与电子技术, 2021, 43(9): 2628-2636.
[3]	阳榴, 朱卫纲, 吕守业, 马爽. 面向非协作多功能雷达的波形单元提取方法[J]. 系统工程与电子技术, 2021, 43(10): 2843-2850.
[4]	季策, 张晓梦. 基于MS-ROMP信号重构算法的欠定盲分离[J]. 系统工程与电子技术, 2020, 42(4): 756-763.
[5]	唐宁, 郭英, 张坤峰. 基于SCA的欠定跳频网台分选方法[J]. 系统工程与电子技术, 2017, 39(12): 2817-2823.
[6]	滕炯华, 徐婧林, 卢隆, 周三平, 韩军伟. 基于自适应参数回归的非局部图像滤波算法[J]. 系统工程与电子技术, 2015, 37(2): 449-454.
[7]	付卫红，李爱丽，马丽芬，黄坤，严新. 基于估计参数势函数法的欠定盲分离[J]. 系统工程与电子技术, 2014, 36(4): 619-623.
[8]	毕晓君, 宫汝江. 基于混合聚类和网格密度的欠定盲矩阵估计[J]. Journal of Systems Engineering and Electronics, 2012, 34(3): 614-618.
[9]	孙燕，尚军亮，刘三阳. 基于采样优化的蒙特卡罗移动节点定位算法[J]. Journal of Systems Engineering and Electronics, 2010, 32(9): 2001-2004.
[10]	曾华，吴耀华，黄顺亮. 非均匀类簇密度聚类的多粒度自学习算法[J]. Journal of Systems Engineering and Electronics, 2010, 32(8): 1760-1765.
[11]	彭耿, 黄知涛, 王丰华, 姜文利. 基于曲线拟合的卫星通信信号参数盲估计[J]. 系统工程与电子技术, 2010, 32(3): 450-453.
[12]	王祖良, 王明昳, 王荔斌, 郑林华. 一种基于信噪比估计的自适应伪码捕获算法[J]. Journal of Systems Engineering and Electronics, 2009, 31(6): 1307-1310.