PBUS: Efficient User Selection for Block Diagonalization in Dense Wireless Networks

Abstract

To mitigate the co-channel interference across adjacent access points (APs) and break the performance bottleneck in dense wireless networks, multiple APs are expected to share the channel state information and cooperatively process the user signals. While the inter-user interference can be eliminated using Block Diagonalization (BD) precoding technique, the number of simultaneous users is limited by the total number of transmit antennas. In this paper, we propose a novel pairing-and-binary-tree-based user selection algorithm (PBUS) to address the user selection issue for multi-user MIMO in dense environments. With the fitness metric evaluated for each pair of users, a binary tree is constructed to store multiple candidate user groups. The best user group is then selected from these candidates. PBUS can achieve both good sum-rate performance and low computational complexity, and also has the flexibility to trade off sum-rate performance and computational complexity.