Low Complexity Spatial Covariance Matrix Reconstruction Algorithm in 2D Hybrid Massive MIMO Systems

Abstract

Massive multiple-input multiple-output (MIMO) has been used in fifth generation (5G) cellular communications. Hybrid analog-to-digital structures offer significant advantages for millimeter-wave communication, such as the reduction in the number of radio frequency (RF) chains. In this structure, the signal at each antenna is not directly received by the digital receiver. Consequently, the spatial covariance matrix, which is an integral part of the direction of arrival (DOA) estimation, is unavailable. In previous studies, the beam sweeping algorithm (BSA) has been found effective for reconstructing the spatial covariance matrix and realizing DOA estimation. However, it has been proven to be computationally intractable. To address this problem and improve the DOA estimation performance, a high-efficiency BSA algorithm aimed at two-dimensional DOA estimation (2D HeBSA) is proposed. The real-valued spatial covariance matrix can be reconstructed by adjusting weights connected to antennas via low-dimensional matrix multiplication. Therefore, super-resolution DOA estimation can be performed using multiple signal classification (MUSIC). Finally, simulations are performed to evaluate our algorithm. According to simulation results, the estimation performance of the proposed algorithm can achieve a better DOA estimation result compared with that of BSA at a considerably lower computational cost.