Research On Random Beamforming Based On Feedback Threshold

Multiple-input multiple-output (MIMO) technique is widely considered as an effective approach to improve the spectrum efficiency, which has become more and more important due to the rapid growth of wireless communication. In the multi-user MIMO (MU-MIMO) system, the throughput of the system can be improved via multi-user diversity (MUD). Random beamforming (RBF) is a precoding technique used to get the MUD, and it has some advantages, such as low-complexity, only need of partial channel state information (CSI) and the same sum-rates scaling law as the optimal Dirty Paper Coding (DPC) with the number of users going to infinity. However, the RBF algorithm also has obvious disadvantage, that is, the excellent performance can be achieved only in the condition of large number of users, meanwhile, each user is asked to feed back individual information to the base station, which makes the uplink channel overloaded.This thesis focuses on the issue that the number of feedback users is too many for the RBF algorithm, which is an onerous requirement on the feedback channel. For two typical communication scenarios, this thesis is devoted to the method for obtaining maximum sum-rates and efficiently reducing the feedback load by setting the feedback threshold to control the number of feedback users in the RBF technique. The main works and contributions are listed as follows.1) For single cell scenario with the inter-beam interference, a novel random beamforming algorithm based on clustering and feedback threshold is proposed. Firstly, the users are divided into multiple clusters according to their signal to noise ratio (SNR). And then, the corresponding feedback threshold in each cluster is calculated by the tool of Extreme Value Theory. Finally, the new RBF algorithm is constructed based on clustering and multiple thresholds. Some theoretical analyses are done on the algorithm. Simulation results not only verify the validity of theoretical analysis but also show that, when compared with the classical RBF technique, the proposed algorithm can considerably reduce the feedback load with a good sum-rates performance.2) The random beamforming with the inter-cell interference is investigated. The random beamforming for multi-cell scenario with homogeneous users is studied firstly. Through theoretical analysis, the homogeneous feedback threshold (HFT) is given with the feedback constraint, and a novel HFT-MRBF algorithm is proposed. The numerical results show that the proposed algorithm has considerably low feedback load and almost no loss of sum-rates. And then, by aid of simulation, the performance of the HFT-MRBF is checked in the context of heterogeneous users. Simulation results show that the proposed method can work well if the fading channels of the users are not obviously different. However, the applicability of HFT-MRBF tends to become worse or even invalid as the difference of users channels increases.