Blockwise-Lattice-Reduction-Aided Tomlinson–Harashima Precoder Designs for MU-MIMO Downlink Communications With Clusters of Correlated Users

This paper presents a new class of precoder designs, namely, blockwise-lattice-reduction-aided Tomlinson-Harashima precoders (BLR-aided THPs), for multiuser multiple-input-multiple-output (MU-MIMO) downlink communications in the presence of clusters (CLs) of correlated channels. By combining the techniques of leakage-based preprocessing and lattice reduction (LR), the proposed BLR-aided THPs decompose the computationally intensive LR procedure into multiple subprocedures, resulting in reduced complexity. A number of precoder designs along with a computationally efficient intercluster ordering algorithm have been proposed. Complexity analysis and simulation results show that the proposed BLR-aided THP designs stand out as useful alternatives in balancing the tradeoffs between complexity and performance in the considered scenario.