基于盖尔圆准则的信源数目估计改进算法  

An Improved Source Number Estimation Algorithm Based on Geschgorin Disk Estimator Criterion

作　　者：褚鼎立 陈红[1] 蔡晓霞 CHU Dingli;CHEN Hong;CAI Xiaoxia(Electronic Countermeasure Institute of National University of Defense Technology,Hefei 230037,China)

机构地区：[1]国防科技大学电子对抗学院,安徽合肥230037

出　　处：《探测与控制学报》2018年第4期109-115,共7页Journal of Detection & Control

摘　　要：针对最小信息准则(Akaike Information Criterion,AIC)存在的非渐进一致性估计的缺陷,以及盖尔圆准则(Gerschgorin Disk Estimator,GDE)可能出现无序特征值导致检测错误的问题,提出了一种基于盖尔圆准则和最小信息准则的GDE-AIC信源数目估计算法。该算法利用盖尔圆半径与噪声模型无关的特性构造似然函数,将其引入AIC准则模型中,克服了AIC准则非渐进一致性估计的缺点,且适用于空间色噪声的环境。在仿真实验中,将该算法与AIC算法及GDE算法等进行对比,结果表明,该方法稳定性好,适用于白噪声与色噪声,且在低信噪比时仍具有良好的估计性能。Aiming at the defect that non-asymptotic consistency estimation of Akaike Information Criterion(AIC)exists and the Gerschgorin Disk Estimator(GDE)may lead to detection errors caused by disorder eigenvalues,a GDE-AIC source number estimation algorithm based on Gerschgorin Disk Estimator(GDE)and Minimum Information Criterion(AIC)was proposed.This algorithm constructed the likelihood function because the Gail circle radius was independent of the noise model to,which was introduced into the AIC criterion model and overcomes the shortcomings of AIC criterion non-gradual consistency estimation for space color noise environment.In the simulation experiment,the algorithm was compared with AIC algorithm and GDE algorithm.Simulation results showed that the proposed method was stable and suitable for white noise and color noise in comparison with AIC algorithm and GDE algorithm,which still had good estimation performance at low SNR.

关 键 词：盲源分离 信源数目估计 最小信息准则 盖尔圆准则 

分 类 号：TN911.7[电子电信—通信与信息系统]