Performance Analysis of MU-MIMO Systems with Threshold-based Feedback and Spatial Heterogeneity

In multiuser multiple-input multiple-output (MU-MIMO) downlink broadcast channel, several users can be simultaneously served through space division multiple access (SDMA) which exploits the multiplexing gain of MIMO systems. To improve the system sum rate, users feedback their channel state information (CSI) to help the base station (BS) schedule those with the best channel conditions. This incurs a large aggregate feedback bandwidth growing linearly with the number of users. There are already several studies investigating methods to reduce the amount of feedback. However, they are limited to spatially homogeneous setting. In this paper, we develop a realistic and general model for the MU-MIMO systems with threshold-based opportunistic feedback, where users are uniformly distributed in the cell (in contrast to the homogeneous setting in which users are of the same distance from the base station). With this model, we analyze the sum rate growth of the system in both the regime with a large number of users and the high signal-to-noise ratio (SNR) regime, and obtain closed form expressions. Moreover, we consider the fairness issue by combining the threshold-based feedback scheme with cumulative distribution function (CDF) scheduling. This proposed method provides fair channel access time ratio (CAR) and tractable analysis.