Fixed-complexity vector perturbation with Block diagonalization for MU-MIMO systems

Abstract

Block diagonalization (BD) is an attractive technique that transforms the multi-user multiple-input multiple-output (MU-MIMO) channel into parallel single-user MIMO (SU-MIMO) channels with zero inter-user interference (IUI). In this paper, we combine the BD technique with two deterministic vector perturbation (VP) algorithms that reduce the transmit power in MU-MIMO systems with linear precoding. These techniques are the fixed-complexity sphere encoder (FSE) and the QR-decomposition with M-algorithm encoder (QRDM-E). In contrast to the conventional BD VP technique, which is based on the sphere encoder (SE), the proposed techniques have fixed complexity and a tradeoff between performance and complexity can be achieved by controlling the size of the set of candidates for the perturbation vector. Simulation results and analysis demonstrate the properness of the proposed techniques for the next generation mobile communications systems which are latency and computational complexity limited. In MU-MIMO system with 4 users each equipped with 2 receive antennas, simulation results show that the proposed BD-FSE and BD-QRDM-E outperforms the conventional BD-THP (Tomlinson Harashima precoding) by 5.5 and 7.4dB, respectively, at a target BER of 10^{-4}.