Study On Adaptive Beamforming In Massive MIMO System

Massive MIMO is one of the key technologies of the fifth generation mobile communication.By deploying a large number of antennas at the transmitter and receiver,the performance and capacity of the system are greatly improved.Because of the large number of antennas in the system,the traditional digital beamforming technology is not suitable for Massive MIMO systems due to hardware cost and other reasons,while low-cost analog beamforming can not meet the performance requirements.Therefore,a digital-analog hybrid beamforming technology,which combines analog beamforming with digital beamforming,has become one of the hot issues in current research.In hybrid beamforming technology,according to whether each RF link is connected to all antennas or only to some antennas,it is divided into full-connection structure and partialconnection structure.Aiming at the downlink scenario of full-connected MIMO system and single-user point-to-point communication,this paper firstly introduces the widely used beamforming method based on orthogonal matching pursuit algorithm and a beamforming method based on singular value decomposition,which can achieve the nearest ideal digital beamforming matrix.Based on the idea of minimizing residual matrix and alternating iteration,a fast convergent hybrid beamforming algorithm is proposed.When designing column vectors of beamforming matrix in analog domain,the modulus of singular vectors of residual matrix is normalized as a candidate set of beamforming vectors in analog domain,and then the beamforming vectors in analog domain are updated according to the singular values.The simulation results show that the proposed algorithm has better spectral efficiency than the orthogonal matching pursuit algorithm in millimeter-wave channel;compared with the column-by-column update algorithm based on singular value decomposition,the spectral efficiency performance is comparable,but it needs less singular value decomposition.For Massive MIMO system with partial connection structure and downlink scenario of single-user point-to-point communication,an improved partial connection structure is proposed.Based on this structure,a hybrid beamforming design scheme with non-constant mode parameters in analog domain is proposed.In traditional analog beamforming/merging,each antenna is implemented by a phase shifter.The optimization problem is a non-convex problem due to the constraints of constant mode in the solution process of analog beamforming,which limits the performance of beamforming.To solve this problem,an improved partial connection structure is proposed.Two parallel phase shifters are equipped for each transmitting antenna.The beam forming parameters in analog domain are realized by adding two unit vectors.Computer simulation results in millimeter-wave channel verify the performance improvement of the proposed scheme.