User selection based on feedback threshold for MIMO broadcast channels

For multi-user multiple-input multiple-output (MIMO) downlink channels, the base station (BS) requires all channel state information (CSI) to select the best user set to achieve multiuser diversity gain. Unfortunately, CSI at the BS have to be fed back from all users for most of systems, especially for the frequency division duplexed (FDD) systems, which imposes a heavy and demanding burden on the feedback channels. Furthermore, more user channel coefficient can result in higher computational complexity for scheduling at the BS. Motivated by these problems, we proposed an improved user selection scheme which sets feedback threshold value based on some existent user selection algorithms. Numerical studies show that the proposed scheme reduces numerous feedback load and computational complexity. However, setting feedback threshold may also incur some loss in sum rate. While simulation results show that if only a proper threshold value is considered the proposed scheme can obtain significant benefits in feedback load and complexity at a cost of little loss in sum rate.