MUSIC算法的DoA估计时间分析  

Analysis of DoA estimation in MUSIC algorithm

作　　者：孙大佳 刘江明[1] 丁良辉[1] 

机构地区：[1]上海交通大学电子工程系,上海200240

出　　处：《信息技术》2015年第12期1-3,9,共4页Information Technology

基　　金：国家自然科学基金(61301117;61221001;61420106008);上海市数字媒体处理与传输重点实验室(12DZ2272600);上海交通大学"科技创新专项资金"(AF0300021)

摘　　要：在基于波束赋形(Beamforming)的无线通信系统中,使用MUSIC算法的波达方向(Direction of Arrival,DOA)估计具有良好的性能。但是,对于MUSIC算法在低信噪比下Do A估计时间和精度之间的关系,却一直未有深入研究。因此,对均匀圆阵数学模型和MUSIC算法进行分析,研究样本数量与DOA估计精度的关系。仿真结果表明,在低信噪比下,DOA估计性能下降,通过增加估计协方差矩阵的样本数量,可以提高MUSIC算法的性能。根据仿真结果,拟合发现DOA估计角度分辨率达到1度,所需样本数量和信噪比的关系服从负指数分布。DoA estimation based on MUSIC algorithm performs well in the communication system using beamforming antennas. However, the research on the relationship between estimation time and accuracy under low SNR is not enough. By analyzing the mathematical model of uniform circular array and deriving the MUSIC algorithm, the relation between sample number and DoA estimation accuracy is proposed. The simulation proves that the performance is degraded under the low SNR. By increasing sample number to estimate the covariance matrix, the performance of MUSIC algorithm is improved. According to the simulation, it finds that the relation between number of samples and SNR follows negative exponential distribution.

关 键 词：DOA估计 MUSIC算法 低信噪比 样本数量 

分 类 号：TN911.7[电子电信—通信与信息系统]