基于Hadamard积的多重镜像压缩传播算子算法研究

Hadamard Product-Based Research on Propagator Algorithm of Multiple Image Compression

传统的传播算子(PM)算法利用矩阵的线性运算代替特征值分解(EVD)得到噪声子空间,在一定程度上降低了运算量,但在整个空间谱的遍历搜索仍需较大计算量,且在低信噪比的情况下估计性能较差.因此,针对空域的一维信号提出了多重镜像压缩的传播算子(MIC-PM)算法,将整个空域等间距地分为多个切片,将原始噪声子空间通过Hadamard积变换,使之从一个切片镜像映射到另一个切片,得到映射噪声子空间,通过映射得到的噪声子空间与导向矢量的正交性得到真实DOA和镜像DOA.理论分析和仿真实验证明,MIC-PM算法能够将PM算法的运算量大大降低,在低信噪比情况下估计精度有所提高.

Traditional propagator algorithm using the matrix of linear operation to replace get noise subspace eigenvalue decomposition（EVD）,to a certain extent,reduces the computational complexity,but the traversal search in the whole space spectrum still needs large amount of calculation,and,in the case of low SNR,the estimation performance is relatively poor.Therefore,in view of the airspace of one-dimension signal,the propagator of multiple image compression algorithm is proposed,and the new algorithm divides the whole airspace into multiple slices.The original noise subspace is mirrored by means of Hadamard transformation from a biopsy mirror mapping to another.By mapping the noise subspace and the orthogonality of the steering vector,true DOA and mirror DOA are obtained.Theoretical analysis and simulation experiments show that MIC-PM algorithm can greatly reduce PM algorithm computational complexity,and,at the same time,under the condition of low SNR,the estimation precision is improved.