中文摘要：

由于在尖峰非正态噪声背景下,MUSIC算法对信源的3D-DOA估计将失去韧性,为此对MUSIC算法进行了改进,提出了一种信源3D-DOA估计新算法。首先简要分析了圆阵模型以及共变,然后将阵列输出矩阵从二阶原点矩扩展到低阶矩,通过分析共变矩阵,得到基于共变矩阵的空间谱,再对空间谱作梯度运算后进行极值搜索,最后可得信源的3D-DOA估计。通过仿真,在尖峰非正态噪声下,算法可以对多个信源的3D-DOA进行正确估计。算法性能分析表明：在背景噪声有冲激且信噪比较低时,FLOC-MUSIC算法较经典MUSIC算法具有良好的韧性。

英文摘要：

Under the spike of non-normal background noise,the MUSIC algorithm for 3D-DOA could lose its capability,therefore this paper improved the MUSIC algorithm,it proposed a new algorithm for 3D-DOA estimation. First it briefly analysed the circular array model and covariation,then extended the moments from two order to low order,through the analysis of covariant matrix,got a spatial spectrum based on covariant matrix,then calculated the gradient of spatial spectrum and searched the extreme,finally got the 3D-DOA estimation. Through the simulation the algorithm could correctly estimate multiple source under the spike of non-normal background noise. The analysis of the algorithm performance show that when the background noise contains impulse and low SNR the FLOC-MUSIC algorithm is better than the classical MUSIC algorithm.