Sparse reconstruction in frequency domain and DOA estimation for one-dimensional wideband signals

Abstract

Previous recovery methods in the literature are usually based on grid partition, which will bring about some perturbation to the eventual result. In the paper, a novel idea for one-dimensional wideband signals by sparse reconstruction in frequency domain is put forward. Firstly, Discrete Fourier Transformation (DFT) is performed on the received data. Then the data of the frequency with the most power is expressed by Fourier serious coefficients. On this basis, the optimization functions and corresponding dual problems are solved. After that the support set is calculated, and the primary sources of this frequency and direction of arrival (DOA) can also be acquired. Comparing with the traditional methods, the proposed approach has further improved the estimation accuracy.