User Selection for MIMO Broadcast Channel with Sequential Water-Billing

摘要

In this paper, we propose a new suboptimal user selection method for the MIMO broadcast channel based on the sequential water-filling (SWF) algorithm. This algorithm is designed to maximize the sum rate of the zero-forcing beamforming (ZFBF) strategy. The reduction of complexity is achieved in a number of ways. First, an iterative procedure to calculate the Penrose-Moore inverse of the channel matrix, based on an LQ decomposition, is introduced. This procedure converts the matrix inversion operation to one vector-matrix multiplication. Secondly, the sequential nature of the user selection problem rules out the try-and-test phase for the water-filling approach. Finally, it is shown that the search space of this algorithm can be further pruned. Simulation demonstrates that this algorithm achieves a sum rate that is close to the maximal sum rate achievable by ZFBF, with a complexity that is comparable to the best algorithms known so far.