压缩感知处理波达方向(DOA)估计问题中,阵列流型的构建是后续估计的基础。该文首先对阵列流型相邻导向矢量的正交性进行理论分析,分析表明在法线方向,等角划分优于等弦划分,在端射方向,等弦划分优于等角划分,相应的DOA估计性能更优。然后,系统推导出等弦划分与等角划分的临界值,并讨论阵元数、划分份数对正交性的影响,设计了一种优化稀疏划分模型,并提出了一种基于等弦和等角空间稀疏相结合的稀疏空域融合(SFSD)DOA估计算法。该算法较等弦划分和等角划分,具有更好的DOA估计稳健性、更低的信噪比门限和更高的估计精度。最后,通过仿真验证了模型的优越性和算法的高效性。

In the estimation of direction-of-arrival (DOA) based on compressive sensing, the construction of manifold matrix is the foundation of subsequent estimates. Firstly, this paper proposes the theoretical analysis about orthogonality of adjacent vector in manifold matrix. The results about the analysis show that the manifold matrix with equal angle interval is superior to that with equal sine interval in normal direction, and the manifold matrix with equal sine interval is superior to the one with equal angle interval in beam direction in the end. Respectively they both perform a better DOA estimation. Then this paper deduces systematically critical value of equal sine interval and equal angle interval, and discusses the influence of the number of array element and division number to the orthogonality. On this basis, a better sparsity model is designed, and a DOA estimation algorithm named sparse fusion in spatial domain (SFSD) is proposed based on the combination of equal sine interval and equal angle interval. Compared with the two other DOA estimation algorithms, the proposed algorithm has the better DOA estimation robustness, lower signal-to-noise ratio threshold and higher estimation precision. Finally, the superiority of the theoretical model and the efficiency of the algorithm are validated by computer simulation.

TN911.7