摘要
MUSIC(Multiple Signal Classification)算法是波达角(the Direction of Arrival,DOA)估计的经典算法之一,但其在二维DOA估计中因需进行二维谱峰搜索而计算量十分巨大.为降低MUSIC算法的计算量,本文在引入变换域DOA概念的基础上提出了一种能够适用于任意阵列结构的二维DOA快速估计算法,即变换域MUSIC(transformed do-main-MUSIC,TD-MUSIC)算法.理论分析和仿真实验表明:该算法不但将空间谱峰搜索的范围减小一半而且具有更低维度的噪声子空间,因而其计算量远小于MUSIC算法.同时,新算法具有比MUSIC更高的空间分辨率.
As one of the most popular techniques for direction-of-arrival(DOA)estimation,the multiple signal classification(MUSIC)algorithm has a tremendous computational complexity for 2-D DOA estimation problems,which is mainly caused by an involved 2-D spectral search step.To reduce the complexity,a new algorithm named the transformed domain MUSIC(TD-MUSIC)for fast 2-D DOA estimation with arbitrary arrays is presented by this work based on a new concept of the transformed domain DOA.It is shown by theoretical analysis and experiment results that TD-MUSIC not only realizes a compression for the spectral search region,but also has a dimension-reduced noise subspace and hence,it requires much lower complexity than the standard MUSIC.Moreover,the new method also shows an improved resolution as compared to MUSIC.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2013年第5期936-942,共7页
Acta Electronica Sinica
关键词
变换域MUSIC
DOA估计
虚拟辐射源
交替投影算法
奇异值分解
transformed domain MUSIC
direction-of-arrival estimation
virtual source
alternate projection method(APM)
singular value decomposition(SVD)