QoS-based user scheduling for multi-user MIMO systems

Next generation wireless networks are being developed to support a wide variety of data services with a broad range of Quality of Service (QoS) requirements. We consider the problem of downlink scheduling wherein a base station (BS) with multiple antennas serves a multiuser system. Assuming that different users can be served by the transmitter antennas at each slot, we allocate each BS antenna to users based on certain priority functions. We study a QoS-based allocation scheme, wherein the priority functions capture the user QoS demands quantified in terms of throughput and delay. We attempt to identify scheduling schemes that provide a good trade-off between (i) higher system capacity by exploiting inherent multiuser diversity, (ii) fairness among users based on their instantaneous channel conditions relative to average channel conditions, and (iii) tolerable latency requirements specified by the user applications. Simulation results evaluate the performance of the above allocation scheme and show that multiple antennas and delay-aware scheduling rules can be efficiently used to provide QoS at little expense of throughput.