Cluster Information based User Scheduling for Multiuser MIMO Systems

Two efficient user scheduling metrics and their corresponding algorithms are proposed for uplink multiuser MIMO systems in a finite scattering environment, where users typically transmit signals while sharing some clusters (which can be trees and buildings). By utilizing statistical channel state information such as how users employ clusters (e.g., Percentage of energy carried by each cluster) and spatial correlation matrices of channels via each cluster, two scheduling metrics which achieve a trade off between decreasing intra-link correlation and increasing inter-link correlation are proposed. Two algorithms based on the proposed metrics are further designed. The proposed algorithms are of low complexity and can be executed in a low frequency through utilizing simple operations and statistical channel information, which brings in a great computational complexity reduction. Simulation results show that both algorithms could achieve significant sum rate gains comparing with other existing algorithms, and the performance can be close to that of the optimal scheme.