User Schedulingalgorithms In Downlink MU-MIMO System

Multi-user MIMO (MU-MIMO) system has multiple antennas at the base station for sending and receiving data. It can transfer data to multiple users or receive data from multiple users simultaneously. In downlink multi-user MIMO system, in order to transmit multiple users simultaneously, pre-coding processing of the data streams needs to be used to eliminate the interference between users in the base station. When the number of antennas is fixed, the number of users that the base station can service in the same time is limited. Then we need to schedule some users to communicate. This paper focuses on the user scheduling algorithms in downlink MU-MIMO system.First we introduced the user scheduling algorithms based on nonlinear TH pre-coding. Based on the classic algorithms classic exhaustive scheduling algorithm and greedy scheduling algorithm, the paper focuses on exhaustive scheduling algorithm based on multi-cycle, greedy scheduling algorithm based on GS orthogonalization and a scheduling algorithm based on channel gain and channel correlation. The simulation results show that:The improved exhaustive scheduling algorithm and improved greedy scheduling algorithm reduce the complexities of the classic algorithms and the system capacities are the same to them. The complexity of the scheduling algorithm based on channel gain and channel correlation is smallest but the system capacity is reduced too.Secondly the paper studies user scheduling algorithm based on linear BD pre-coding. Based on classic exhaustive scheduling algorithm and greedy scheduling algorithm, it focuses on two improved scheduling algorithms: the Frobenius norm-based scheduling algorithm and scheduling algorithms based on matrix volume. A low-complexity user scheduling algorithm based on the angle between the users is proposed on the basis of conventional algorithm and greedy algorithm. The simulation results show that, the total system capacity of the proposed algorithm is close to that of greedy scheduling algorithm. In the case of high SNR, the performance is better than the Frobenius norm-based scheduling algorithm and scheduling algorithms based on matrix volume. In order to ensure fairness in some systems, we improve the existed algorithms by processing a round-robin (RR) user selection. The scheme divides the total service time into several time slots.In each time slot recursively apply the user scheduling algorithms to select the users. All users in the system have the opportunity to be served by the base station, so the scheme can ensure the fairness.Finally the paper studies two codebook-based limited feedback schemes combined with user scheduling: channel vector quantization (CVQ) and per-user unitary rate control (PU2RC). The simulation results show that: The total system capacity of the CVQ scheme is larger than the PU2RC scheme when the number of scheduled users is limited.