Low-Complexity Lattice Reduction-Aided Channel Inversion Methods for Large-Dimensional Multi-User MIMO Systems

Abstract: Low-complexity precoding {algorithms} are proposed in this work to reduce the computational complexity and improve the performance of regularized block diagonalization (RBD) {based} precoding {schemes} for large multi-user {MIMO} (MU-MIMO) systems. The proposed algorithms are based on a channel inversion technique, QR decompositions{,} and lattice reductions to decouple the MU-MIMO channel into equivalent SU-MIMO channels. Simulation results show that the proposed precoding algorithms can achieve almost the same sum-rate performance as RBD precoding, substantial bit error rate (BER) performance gains{,} and a simplified receiver structure, while requiring a lower complexity.