基于IMF自相关相位矩阵的改进型信源个数估计算法  

Improved source number estimation algorithm based on IMF autocorrelation phase matrix

作　　者：葛生国 潘晴[1] 田妮莉[1] Everett X.Wang GE Shengguo;PAN Qing;TIAN Nili;Everett X.WANG(Faculty of Information Engineering,Guangdong Universityof Technology,Guangzhou 510006,China)

机构地区：[1]广东工业大学信息工程学院

出　　处：《电视技术》2019年第10期1-5,14,共6页Video Engineering

基　　金：国家自然科学基金项目(61901123)

摘　　要：本文提出了一种基于本征模态函数(Intrinsic Mode Function,IMF)自相关相位矩阵的改进型信源个数估计方法,分别对均匀圆阵(Uniform circular array,UCA)接收的远场窄带信号的各IMF相关相位矩阵做特征分解获得特征值,在此基础上进一步特征提取,获得更高维的特征向量,利用BP(back propagation)神经网络训练新的信源个数估计模型。实验表明:本文方法在小快拍数下具有更优秀的信源个数估计性能。In this paper,an improved method for estimating the number of sources based on the autocorrelation phase matrix of the intrinsic mode function is proposed.The eigenvalues are obtained by decomposing the IMF correlation phase matrices of the far-field narrowband signals received by the uniform circular array.Based on this,further feature extraction is performed to obtain higher dimensional eigenvectors,and BP neural network is used to train a new source number estimation model.Experiments show that the proposed method has better performance of estimating the number of sources under small snapshots.

关 键 词：本征模态函数 均匀圆阵 信源数估计 BP神经网络 小快拍数 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]