一种基于四阶累积量的空间特征盲估计方法

该文以均匀直线阵为基础，提出了一种基于阵列输出四阶累积量的信号空间特征盲估计方法，并将其用于相干多径环境下的多用户信号空间特征的估计。作者利用四阶累积量构造了一个空间特征估计矩阵，并证明了对该矩阵作特征分解可以得到各用户的空间特征。该盲算法不依赖于信号具体特征，适用于任意加性高斯(白或有色)噪声。理论分析和仿真结果说明了算法的有效性和稳健性。

A Blind Spatial Signature Estimation Algorithm Based on 4th Order Cumulant

Based on uniform linear array and 4th order cumulant of array output signals, a novel blind algorithm for spatial signature estimation was presented in this paper, and was applied to estimate spatial signature of multi-user with coherent multi-path. Using 4th order cumulant, the author constructed a spatial signature matrix , and then proved that multi-user’s spatial signature can be obtained through eigen decomposition of the matrix. The algorithm is not depended upon the specific characteristic of signal, and can be employed to additive arbitrary Gaussian (white or colored) noise. Theoretical analysis and computer simulation confirmed the availability and robustness of the algorithm.