Abstract:  For the problem of broadband underdetermined direction of arrival (DOA) estimation, a novel DOA estimation algorithm is proposed based on the continuous sparse recovery. Firstly, the dimension of coprime array receiving data is reduced by using direction wavenumber and the covariance matrix is vectorized to improve the degree of freedom. Then, the continuous sparse recovery model of direction wavenumber is established using the spatial sparseness of direction wavenumber, and the estimation of direction wavenumber is achieved with convex optimization and polynomial rooting. Finally, the signal frequencies and direction wavenumbers are pair matched by using Capon method. With this method, offgrid effects caused by discretizing this range onto a grid in traditional sparse recovery can be neglected, it also improves accuracy and resolution of DOA estimation and the number of sources estimated by the proposed algorithm is larger than the number of actual arrays. Theoretical analysis and simulations demonstrate the effectiveness and feasibility of the proposed method.

Key words:  array signal processing, direction of arrival (DOA), broadband, continuous sparse recovery