说明：  MUSIC算法是一种基于矩阵特征空间分解的方法。从几何角度讲，信号处理的观测空间可以分解为信号子空间和噪声子空间，显然这两个空间是正交的。信号子空间由阵列接收到的数据协方差矩阵中与信号对应的特征向量组成，噪声子空间则由协方差矩阵中所有最小特征值（噪声方差）对应的特征向量组成。
(Music algorithm is a method based on feature space decomposition of matrices. From the point of view, observed spatial signal processing can be broken down as the signal subspace and noise subspace, it is clear that the two spaces are orthogonal.Signal - subspace spatial data received by the array corresponding eigenvector of the Covariance matrix and signal group)