DOA Estimation of Coherent Signals Based on Atomic Norm Minimization in Signal Subspace

Most of the direction of arrival (DOA) coherent signal estimation methods based on coprime array do not consider the influence of noise. Therefore, in this paper, we propose a signal subspace based atomic norm minimization (ANM) method to estimate the DOA of coherent coprime array. Firstly, we use the relationship between the linear array and the coprime array to directly map and interpolate the matrix containing only the signal subspace in the coprime array to obtain the signal subspace of the linear array, and then solve the ANM problem to recover the rank. Finally, the DOA of coherent sources was obtained through the application of the matrix enhancement and matrix pencil (MEMP) method to construct the enhancement matrix. The proposed method can estimate the coherent signal more accurately and is more robust to noise. The results of numerical simulations substantiate the superiority of the proposed algorithm.