改进Geese算法对信号的DOA估计

针对传统的空间谱估计算法对相干源的波达方向（DOA）估计会失效的问题，该文在利用信号子空间特征向量生成广义特征值（Generalized Eigenvalues utilizing Signal subspace Eigenveetors，Geese）算法基础上，结合空间平滑技术对接受数据进行预处理，提出一种改进的Geese算法。该算法由于不需方向搜索且只利用信号子空间，大大降低了计算复杂性。计算机仿真结果表明，与Geese算法相比，该改进算法既能够有效地估计出独立信号源的DOA，也能有效地估计出相干信号源和相隔比较近的低信噪比信号源的DOA。对于独立源的估计，在相同条件下，估计性能也要优于经典的空间谱估计算法（ESPRIT算法和MUSIC算法），从而证明该算法的合理性。

Estimation of Signal'DOA by Modified Geese Algorithm

In this paper, a modified Geese algorithm is proposed which combines spatial smoothing technique and can maintain its computational simplicity. The simulations results have proved that this algorithm is effective for estimating the DOA either of independent signals or of coherent signals and adjacent signals with small SNR and in the same conditions has better performance than standard ESPRIT and MUSIC for independent signals.