Analysis of Multiuser Diversity Gain Achieved by MIMO Systems Through User Scheduling Based on Limited Feedback in Dense Cellular Networks

We examine the sum rate of multiple-input multiple-output broadcast channels in which the locations of base stations (BS) and users follow independent homogeneous Poisson point processes. Each BS schedules appropriate users for spatial multiplexing based on partial channel state information provided by users. In this scenario, we develop a suitable channel quality information (CQI) for user scheduling, which attains a nearly optimal multiuser diversity gain. Using the proposed CQI, we determine the growth rate of the sum rate achieved with limited feedback in densely populated cellular environments. In comparison to the single-cell scenario, the growth rate is slightly diminished due to inter-cell interference caused by the dense BS topology. Nevertheless, we demonstrate that the gap in net sum spectral efficiencies achieved through user scheduling with the proposed CQI remains significantly large, even within dense cellular networks, provided that each BS can employ the optimal number of feedback bits. Thus, we also investigate the optimal feedback rate that maximizes the net sum spectral efficiency.