Off-grid DOA estimation based on noise subspace fitting

Dictionary mismatch caused by finite spatial discretization and off-grid situation leads to performance degradation in sparse-representation-based direction-of-arrival (DOA) estimation. In this paper, a procedure implementing DOA estimation and rectification of dictionary in an alternating way is designed. A strategy using noise subspace fitting (NSF) is proposed to estimate the direction bias and therewith treat the dictionary mismatch. Based on NSF, the dictionary rectification model is extended from first-order to second-order Taylor approximation to achieve higher modeling accuracy. Simulation results show that improved DOA estimation performance can be achieved for off-grid targets.