User Selection In The MU-MIMO System Based On Limited Feedback

Multi-User Multiple Input and Multiple Output(MU-MIMO)techonlogy using multiple antennas in both the user terminal and base stations which greatly improves the spectrum efficiency of the communication system.Since multiple users use the same frequency within one time slot in the MU-MIMO system,the Co-Channel Interference(CCI)among users should be eliminated by using pre-coding in the base station,and the Channel State Information(CSI)is needed to design the pre-code.The base station can obtain the CSI of downlink channel by the information feedback of the user terminal,but under the situation of large number of users and transmission antennas,the limited uplink channel spectrum can't afford large amount of feedback overhead.In order to reduce the amount of feedback,the same codebook need to be preset in both the base station and user terminal.The codebook contains all possible values of the CSI,each value is called code-word.The user terminal only needs to feedback the selected code-word index to the base station to indicate the CSI,thus the feedback overhead is reduced.The code-word will be used to do pre-coding and user selection.The user selection algorithms usually selects users by using the criterion of maximizing the sum of user's achievable bit rate(system capacity).In most literatures,the user selection algorithm is designed in the case that the base station can obtain accurate CSI(_kV)and CCI is eliminated completely.But the quantization error of code-word will lead to residual CCI.Only few literatures have done research on how to select users with residual CSI.In the literatures[16][17],the one-dimensional SINR caused by residual CCI is derived under the circumstances that the base station uses the equivalent channel corresponding to the maximum singular value to transfer data,then the papers derive the system capacity based on the SINR and design the corresponding user scheduling algorithms.But this kind of literatures only uses one equivalent channel which wastes MIMO space channel resources.This is because the channel state of multiple equivalent channels of MIMO are different and it is difficult to give a uniform SINR that reflects user's achievable bit rate,without the SINR,it's difficult to design user selection algorithm based on maximum system capacity.In this paper,the concept of multi-channel integrated SINR is proposed for limited feedback MU-MIMO system where all equivalent channels are used to transmit data.This paper proves that the achievable bit rate of user k depends on the proposed multi-channel integrated SINR.We design user selection algorithm based on the integrated SINR and proves that the computation load of this algorithm is smaller than that of computing user's achievable bit rate to select users directly.And the algorithm designed by us can obtain a larger system capacity compared with algorithms of literatures only using one equivalent channel to transmit data.In summary,our contributions are:1.We propose the concept of the multi-channel integrated SINR and derive the relationship between the multi-channel integrated SINR and the user's achievable bit rate.2.Under the condition of using all equivalent channels to transmit data,we design a user selection algorithm based on multi-channel integrated SINR which can obtain larger system throughput compared with the existing literatures only using one equivalent channel to transmit data.3.The proposed algorithm has a smaller computation load compared with algorithms which directly calculate user's achievable bit rate to select users.