摘要
基于信息论的MDL准则信源数估计方法在均匀高斯白噪声背景下具有良好的估计性能。若阵元间存在幅相不一致性时白噪声变成非均匀色噪声,此时算法性能将急剧下降。针对该问题,提出了基于数据协方差矩阵特征值总体最小二乘线性拟合误差的信源估计方法。该方法根据特征值线性拟合误差与拟合点数为单调不增关系,并结合MDL准则的惩罚因子构造判断函数实现信源数估计。仿真结果表明:该方法在非均匀色噪声背景下能够实现信源数一致估计;且在低信噪比、短快拍时仍然具有良好的估计性能。
The MDL criterion based on information theorem have excellent performance under the condition of uniform Gauss white noises. However,the performance of the MDL method degrades significantly when additive noise assumptions situation against deviations from temporally or spatially white due to channelphase-amplitude imbalance in the array. A source number detection method used in the presence of unknown spatially nonuniform noise is proposed based on the total least squares（TLS） fitting with the eigenvalues. Instead of relying on the eigenvalue distribution in the MDL criterion,the work devises linear fitting error by the method of least squares. Fitting error term will be a monotonically increasing function of fitting number of descending eigenvalue. The criterion for the proposed method is proposed incorporates an adjusting term which takes into account the finite sample size. The proposed criterion gives a consistent estimation of the number of signals in a nonuniform noise environment,especially in the low signal-to-noise ratio and small samples. Numerical examples are provided to demonstrate effectiveness of the proposed approach.
作者
陈明建
龙国庆
黄中瑞
CHEN Ming-jian;LONG Guo-qing;HUANG Zhong-rui(School of Electronic Countermeasure,National University of Defense Technology,Hefei 230037,Chin)
出处
《火力与指挥控制》
CSCD
北大核心
2018年第7期32-36,共5页
Fire Control & Command Control
基金
安徽省自然科学基金资助项目(1608085QF140)
关键词
阵列信号处理
信源数估计
非均匀噪声
总体最小二乘
特征值分解
array signal processing,source number detection
nonuniform noise,total least squares,eigenvalue decomposition
