Optimization of Sparse Planar Arrays Based on Extended Unitary Matrix Pencil Method

In this paper, a hybrid scheme of extended unitary matrix pencil (EUMP) and random singular value decomposition (RSVD) is presented for the synthesis of sparse planar arrays with high efficiency. The pattern matching accuracy is improved owing to unitary transformation, which converts the sample matrix from complex to real along with its eigenvectors. Memory consumption and computation complexity of matrix factorization are reduced significantly by using the RSVD instead of the SVD. Simulations verify that the method can achieve accurate reconstruction of a given pattern effectively.