盖氏圆准则信源数估计算法的分析与改进

Research and Improvement on the Estimator Using Gerschgorin Disks

为了减少盖氏圆准则信源数估计算法的运算量并且提高信源数估计的精度,根据噪声空间与阵列导向矩阵的正交性原理,设计了基于特征空间的信源数估计算法(Estimator Based on Eigenvectors,EBE).EBE构造时空相关矩阵,利用色噪声在时间上相关性比较弱的特点,实现对空间色噪声的抑制.在空间白噪声环境下和空间色噪声环境下测试了EBE信源数估计的性能并且与传统的盖氏圆准则和其它色噪声类信源数估计的一些算法比较,证明了EBE在空间白噪声和空间色噪声环境下的有效性.EBE不仅节省了盖氏圆准则信源数估计中一次特征分解的运算量,并且同时提高了信源数估计的性能.

In order to reduce the computer burden of Gerschgofin radii estimator, according to orthogonal property of noise subspace to array manifold, presents a novel and effective algorithm（ Estimator Based on Eigenvectors, EBE）. EBE exploits the correlation property of colored noise in time to restrain it through constructing a novel time-space correlation matrix. Proves the validity of EBE theoretically and applies EBE successfully in computer simulation. In both environments of while gauss noise and colored noise EBE could estimate the number of incoming waves with less error than the estimator using Gerschgorin disks and some other signals number estimators designed for colored noise. EBE not only save the computer burden of one correlation matrix decomposition in Gerschgorin radii estimator, but also improve the performance of signals number estimation.