Research Of Beamforming In Massive MIMO System

To further meet the people’s urgent requirements of the high quality of mobile communications,massive MIMO,as the key technology for increasing the data transfer rate,has become a hot topic of 5G.However,with the growth in the number of antennas,the complexity of the signal processing is increasing,and the traditional MIMO beamforming is difficult to directly apply to massive MIMO scenes.In particular,due to the presence of high-dimensional matrix inversion,the complexity of the near-optimal regularized zero-forcing(RZF)beamforming is almost unacceptable.This thesis mainly research on low-complexity beamforming technology of massive MIMO.In order to reduce complexity,this thesis research on an optimization algorithm of the RZF beamforming algorithm.At the beginning of this algorithm,the ordinary conjugate symmetric matrix inversion is replaced with a Toeplitz matrix inversion.And the Toeplitz property is based on the approximate orthogonality and symmetry between massive MIMO antennas.Then the Toeplitz matrix is further evolved into a banded Toeplitz matrix,as most of the Teoplitz matrix elements tend to zero.Finally the banded Toeplitz matrix inversion is solved efficiently by the improved Levinson-Durbin algorithm.Further by calculating the number of floating-point operations and the simulation under different channel quality at the end of this thesis,it can be concluded that the per-user rate loss of our algorithm compared to that of RZF is relatively small and accepted with the large complexity reduction.This conclusion proves the effectiveness of the optimization algorithm.