Robust Beamforming for Massive MIMO System with Arbitrary Array Geometry

Digital beamforming method has been widely applied in MIMO systems, but it is not suitable for massive MIMO systems due to the unbearable power consumptions and costs it causes. Hybrid beamforming becomes a cost-effective solution for massive MIMO beamforming, though there is usually trade-off between performance and computation complexity. How to design a hybrid beamforming method to achieve good performance while with a low complexity becomes a challenge. In this paper, we propose such a method based on eigenvalue decomposition and regularized channel diagonalization (EVD-RCD). Without knowing prior knowl-edge of the array geometry, the proposed algorithm not only achieves high spectral efficiency with low complexity, but also has strong robustness to deal with different antenna array geometry, imperfect CSI and quantized phase shifter.