Adaptive DOA Estimation with Low Complexity for Wideband Signals of Massive MIMO Systems

Abstract For massive multiple-input multiple-output (MIMO) communication systems, it is a major challenge to develop direction-of-arrival (DOA) estimation algorithm with low complexity and high accuracy for wideband signals. We propose a low-complexity adaptive multistage Wiener filter (MSWF)-based two-sided correlation transformation (AM-TCT) algorithm for wideband DOA estimation of two-dimensional (2D) massive MIMO systems. Unlike conventional TCT algorithms, the proposed AM-TCT algorithm uses the signal subspace that is obtained by forward recursion of the MSWF to construct the focusing matrix, which can reduce the computational complexity. To further reduce the computational complexity, the eigenvalue decomposition (EVD) of the covariance matrix is also replaced by the MSWF. Furthermore, to improve the precision, the Minimum Description Length (MDL) criterion uses the stage of MSWF to adaptively select the appropriate dimension of the noise subspace, and the backward recursion of the MSWF is employed to accurately estimate the noise subspace. Theoretical analysis demonstrates the complexity superiority of the proposed AM-TCT algorithm. Simulation results indicate that the proposed AM-TCT algorithm can effectively estimate the angle of wideband sources in massive MIMO systems and outperform some existing methods.