Abstract: A method of estimating direction of arrival which is based on Krylov subspace is proposed.Since Krylov subspace is equivalent to the signal subspace under some condition，by forming basis depended on incident angle and testing there equivalence，DOA can be obtained.The algorithm is done through multistage Weiner filtering.Simulation shows that the proposed method offers improvements in estimation performance for closely spaced signals.

Key words: Direction of Arrival（DOA） estimation, Krylov subspace, multistage Weiner filtering

摘要： 提出一种基于Krylov子空间的信号波达方向估计算法，搜索信号可能入射角度，通过测试所构造的Krylov子空间与信号子空间的等价性来判断信号的DOA，算法用多级维纳滤波实现子空间分解。仿真实验表明，算法在低信噪比条件下对相邻信号有良好的谱分辨率和估计性能。

关键词: 波达方向估计, Krylov子空间, 多级维纳滤波