The Application Of Random Beamforming In Mimo Scenarios

With 4G cellular technology now beginning to be deployed widely around the world, the fifth generation(5G) mobile and wireless communication technologies are emerging into research fields. The application of Massive MIMO is receiving much more concerns. Compared with traditional MIMO technology, much more antenna will be used in the base station in Massive MIMO system, and much more throughput gain is shared by the terminals. But due to the huge number of low power base station antennas, the feedback load of Channel State Information(CSI) is large. Choosing a suitable precoding algorithm will greatly impact the performance of Massive MIMO system. The Random Beamforming(RBF) is one of the precoding algorithms which needs less feedback on CSI, and will have huge potential.Within a single-cell system, the application of the RBF has been widely studied. By model building, we know that Multi-user Spatial Multiplexing, and serves a large number of users can be used in the RBF. What’s more, the demand of CSI feedback is low, thus the system also has low complexity advantage. Then this paper analyzes how the performance of RBF can be affected by the number of the beams the base station transmits at one time. And we conclude that by choosing the proper number of the beams according to different channel environments, the system performance can be improved greatly. Then we also come up with Multi beam selection(MBS) mechanism. The core idea of the MBS is that the total power of the base station will be reasonably distributed to each beam, at the same time the energy utilization efficiency is improved, by this way the system performance is improved. Per unitary basis stream user and rate control(PU2RC), which is a generalization of RBF, and we make a comparison between PU2 RC and ZFBF. On the condition of the Multi-MIMO system with the lack of perfect CSI, only the feedback for the SINR of each user and index corresponding to each beam is needed by the PU2 RC, and the complexity is low, and better performance can be achieved. But the disadvantage of this algorithm is that the performance of the PU2 RC will decline sharply with the number of users decrease.On the scenario of mult-cell systems, the system model of the RBF algorithm is introduced to a much deeper degree. RBF is applied at each BS, at the same time, either minimum-mean-square-error(MMSE), matched filter(MF), or antenna selection(AS) based spatial receiver is employed at each mobile terminal(denoted as RBF-MMSE, RBF-MF, and RBF-AS schemes, respectively). By designing the user receivers, we can eliminate both the intra-cell and inter-cell interferences effectively. By MATLAB simulation, we obtain the conclusion that the performance of MMSE is the best, while the complexity is the highest; ZF takes the second place, and is easy to implements also; and AS takes the worst performance, while is the most easy to implement. A variety of methods to validate the conclusions is presented in this paper.In the end of this paper, the content is about the application of precoding algorithms in Massive MIMO, and performance comparison is made between Zero-force(ZF) and matched filter(MF) among the users in center and margin areas. At last, novel mechanisms named kmeans and K-Medoids are presented to group users to implement the hierarchical precoding scheme. By MATLAB simulation, we conclude that the complexity of these two mechanisms is nearly the same, but the K-Medoids grouping mechanism is much better in performance.