摘要
针对传统MUSIC算法在信噪比较低的条件下分辨力下降的问题,提出一种适用于宽带信号的改进MUSIC算法,首先对接收信号进行快速傅里叶变换,将时域信号转移至频域,并划分子带;对每个子带上的信号求解协方差矩阵,进行特征值分解,构造信号子空间和噪声子空间;将信号在两个子空间的投影比值作为参考,进行子带波达方向估计,进而平均求出最终宽带信号波达方向。仿真结果表明,在低信噪比条件下,相比于传统MUSIC算法,所提方法具有较好的空间分辨力。
To solve the problem that the traditional MUSIC algorithm decreases the resolving power when the signal-to-noise ratio is low,an improved MUSIC algorithm for wideband signals is proposed.Firstly a fast Fourier transform is used to transfers the receiving time domain signal to the frequency domain and divide them to subbands.The covariance matrix is solved for the signals in each subband and perform eigenvalue decomposition to construct the signal and noise subspaces.The ratio of the projections of the signal in the two subspaces is used as a reference to estimate the subbands'direction,and then the final broadband signal direction is averaged out.The simulation results show that the MUSIC algorithm has better spatial resolution than the traditional MUSIC algorithm under the low signal-to-noise ratio.
作者
张远驰
胡进
ZHANG Yuanchi;HU Jin(The 724th Research Institute of China State Shipbuilding Corporation,Nanjing 210000,China)
出处
《电声技术》
2023年第10期97-99,共3页
Audio Engineering
关键词
MUSIC算法
快速傅里叶变换
特征值分解
方向估计
MUSIC algorithm
fast Fourier transform
eigenvalue decomposition
direction estimation
