Low Complexity Orthogonal Matching Pursuit Based Channel Estimation in mmWave Massive MIMO Systems

In mmWave channel estimation problem for massive MIMO systems, the compressed sensing (CS) techniques incurs prohibitive computational complexity, especially when the transmitter and receiver are both equipped with large number of antennas. In this paper, we convert the channel estimation problem into the equivalent multidimensional CS problem where the sparse channel is represented by a set of independent dictionaries with lower dimension instead of a large single one. Then we propose a multidimensional orthogonal matching pursuit relying on parameter refinement (MOMP-PR) algorithm to reconstruct the channel with lower complexity. Specifically, it performs the matching projection and updates the residual observation using the MOMP part to obtain the coarse estimation of channel parameters. Subsequently, these parameters can be refined to reduce the estimation error incurred by the basis mismatch between the continuous angles in reality and on-grid angles in the dictionary. Finally, simulation results demonstrate the effectiveness of our proposed algorithm.