基于稀疏表示和约束优化的波达方向估计方法

对于噪声环境中信号源的波达方向(DOA)估计,传统的多信号分类(MUSIC)算法只对不相干信号有效，且所需较多样本。针对此问题，将进行DOA估计的搜索范围看作冗余字典，从而待估计的DOA成为该冗余字典中的某些元素，可以由冗余字典对其进行稀疏表示；其次，利用单次快拍数据，应用二阶锥(SOC)约束优化的方法对该稀疏表示问题进行描述，并进而转化为标准的二阶锥形式，采用有效的优化工具SeDuMi来实现DOA的估计。仿真结果表明，与现有的子空间方法相比，该方法只需单拍数据即可得到较好的估计结果，且无需对信源个数有先验知识，同时适用于相干和非相干信号。

DOA estimation method based on sparse representation and constrained optimization

For Direction-Of-Arrival(DOA) estimation of signal in additive noise,the traditional Multiple Signal Classification(MUSIC) algorithm cannot process the coherent signal with fewer snapshots.The searching scope of estimated DOA was considered as redundant dictionary.Consequently,the estimated DOA was taken as some elements in the dictionary,and could be represented sparsely by the dictionary.Then,this problem was thrown into the Second Order Cone(SOC) constraints and an efficient estimation algorithm using a single snapshot was developed.This constrained problem could be depicted as a standard SOC form and be solved by the SeDuMi,an optimization toolbox.The simulation results show that the proposed algorithm has a few advantages over the existing subspace method including one single snapshot to be needed,no requirement for the number of source signals,ability to work with coherent and non-coherent signals.