MUSIC测向算法的入射方向确定方法

MUSIC(MUltiple SIgnal Classification)等谱峰搜索类算法的计算量主要集中于谱峰搜索过程，加大搜索步长虽可减小计算量，但会引起空间谱的畸变从而造成测向误差。先采用大步长搜索确定谱峰大约位置后再用小步长搜索可以减小测向误差，且增加的计算量较小，但其实时性不佳。文中提出了基于重心估计的入射方向确定方法，利用现有谱峰搜索数据进行后处理，当搜索步长远大于MUSIC测向算法本身的均方根误差(RMSE，Root Mean Square Error)时，可用很小的计算量大大提高测向精度，具有良好的工程应用价值。 

Estimating Directions of Incident Signals in MUSIC Algorithm for Direction-finding

The computation of peak-searching algorithms such as MUSIC (MUltiple SIgnal Classification) is mainly concentrated in the process of peak-searching. One of countermeasures is to increase the searching step. Meanwhile, the estimation errors arise because the spatial spectrum is distorted. A compromise method is to search with a minor step after searching with a larger step to acquire the approximate incident angles. While it is time costly because of its serial process of 2-step searching angles. An angle estimation method based on the centrobaric estimation for MUSIC algorithm is proposed in this paper. It only needs post-processing upon the peak-searching data the same as before. When the searching step is much larger than the essential RMSE (Root Mean Square Error) of MSUIC, the presented method can greatly improve the accuracy by a tittle computation. It is useful to practical engineering applications.