A study on off-grid issue in DOA and frequency estimations

Abstract

This work addresses the off-grid issue for DOA and frequency estimations when the dictionary based sparse signal recovery concept is adopted. By off-grid, we mean that the true values of signal, angles or frequencies in this case, are not exactly on the sampling grid created by utilizing the discrete dictionary technique. To handle this problem, off-grid is remodelled such that it is represented by an offset matrix that is a sparse matrix. And then, a direct estimate of the offset matrix is developed to compensate the off-grid by utilizing the fact that the offset matrix is a sparse matrix. Finally, by exploring the sparse property of DOAs/frequencies and offset matrix, a joint estimation approach is devised under optimization framework. Numerical studies demonstrate the effectiveness of the proposed approach.