Reduced Complexity Direction-of-Arrival Estimation for 2D Planar Massive Arrays: A Separation Approach

This paper proposes a low complexity direction of arrival (DOA) estimation method for two-dimensional (2D) massive multiple-input multiple-output (MIMO) systems. The conventional Multiple SIgnal Classification (MUSIC) algorithm of an example of 2D DOA estimation has a great DOA estimation performance, although there is a problem that the calculation amount becomes enormous due to 2D spectral search, especially in massive MIMO systems. By using a part of row or column antenna elements, it is possible to separate search domains azimuth and elevation, respectively. Observation dimension can be degenerated to 1D and subset of row/column-wise antenna elements are sufficient for DOA estimation. The proposed approach can reduce computation complexity while achieving an improved DOA estimation accuracy. Simulation results validate its effectiveness and superiority. The obtained findings will be useful for designing practical massive MIMO systems in future mobile communications.