Projection Riemannian Manifold based Regular Sparse Array Beamforming for Millimeter Wave Communication

In the Fifth Generation (5G) Millimeter-Wave (mmWave) networks, large scale Uniform Rectangular Array (URA) with high directive beamforming gain is utilized to achieve the coverage requirements. However, the massive antennas of large-scale URA generally brings high hardware cost and power consumption. In this paper. Projection Riemannian Manifold (PRM) based Regular Sparse Array (RSA) beamforming is proposed to reduce hardware cost and power consumption. Firstly, RSA obtained by removing parts of antenna elements from a uniform array is introduced into mmWave networks, which enables implementation feasibility. The mathematical modeling and residual error algorithm is proposed to derive minimal antenna number for RSA. Furthermore, PRM is proposed to generate beam book with non-linear constraint, aiming to boost the coverage with less antennas. In addition, we’re the first to propose P-R beamforming framework interlaced both PRM and RSA with MMSE verification. From performance evaluation and theoretical analysis, the proposed P-R beamforming algorithm can reduce at least 15°/o hardware cost and power consumption with negligible performance loss.