Skip to Main Content
In this paper, we propose a low-complexity fair scheduling algorithm for wireless multiuser MIMO communication systems in which users are multiplexed via time-, frequency-, and space-division multiple access (SDMA) schemes. In such systems, the transmission quality considerably degrades if users with spatially correlated channels are to be served at the same time and frequency. The approach presented here works with both zero- and nonzero-forcing SDMA precoding schemes by deciding, for each time and frequency slot, which users are to be served in order to maximize the precoding performance. The number of users is not a fixed parameter of the algorithm (as often assumed for other schedulers present in the literature), but it is also adjusted in accordance to the channel conditions. While smaller SDMA groups allow us to transmit with a higher average power per user, larger groups lead to higher multiplexing gains. Our algorithm ProSched is based on a novel interpretation of the precoding process using orthogonal projections which permit us to estimate the precoding results of all user combinations of interest with significantly reduced complexity. In addition, the possible user combinations are efficiently treated with the help of a tree-based sorting algorithm. The ProSched takes advantage of a perfect channel state information, when available, or, alternatively, of second-order channel statistics. The individual-user quality-of-service requirements can be considered in the decision-making process. The effectiveness of the algorithm is illustrated with simulations based on the IlmProp channel model, which features realistic correlation in space, time, and frequency.