多级维纳滤波是一种降维算法，文中将多级维纳滤波引入到多重信号分类（MUSIC）算法中，提出了适合于MUSIC方法的多级维纳滤波器结构。该方法避免了采样数据协方差矩阵的特征值分解，降低了运算量。仿真结果证明了方法的有效性。

Multi-stage Wiener filter is a reduced-rand algorithm. In this paperw the multi-stage Wiener filter is introduced into MUSIC algorithm. The suitable configuration of multi-stage Wiener filter for MUSIC is proposed. The new method has less computational cost for the new method need not the eigenvalues decomposition. Simulation results demonstrate the effectiveness of the method.

TN957

国家自然科学基金资助项目（60502040）