基于Krylov子空间的宽带信号DOA快速估计方法

提出了一种基于Krylov子空间的宽带信号DOA（Direction of Arrival）快速估计方法。该方法首先对方向矩阵进行Jacobi—Anger展开来构造“聚焦”矩阵，然后通过多级维纳滤波（MSWF：Multi-Stage Weiner Filter）算法求得聚焦后的阵列协方差矩阵的Krylov子空间，因为在满足一定的条件下，Krylov子空间等价于阵列的信号子空间，所以可以求得信号的DOA。采用Jacobi-Anger展开式构造聚焦矩阵不需要进行角度的预估计，通过MSWF算法求Krylov子空间不需对观测数据的协方差矩阵进行特征值或奇异值分解，从而使得该方法运算量比常用的宽带空间谱估计方法要小。计算机仿真试验证实了方法的有效性。

A Fast DOA Estimating Method of Wideband Signals Based on Krylov Subspace

A fast DOA estimation method of wideband signals based on Krylov subspace has been proposed in the paper. At first, A focusing matrix is formed by expanding the steering matrix using the Jacobi-Anger expansion, then the Krylov subspace is educed using the Multistage Weiner Filtering arithmetic. Because the Krylov subspace is equivalent to the signal subspace under some conditions, signals DOA can be estimated. This method needs no direction pre-estimation or eigenvalues decomposition, which needs less computational cost than other methods. Simulation results demonstrale the effectivenes of the method.