Precoding Design for Multiuser MIMO Downlink Communication Systems Using an LQ-Based Cascade GSIC Algorithm

This paper proposes a novel LQ-based generalized side-information cancellation (GSIC) precoder for multiuser multi-input multi-output (MU-MIMO) downlink systems with sum-rate performance enhancement. The proposed transceiver comprises the following stages. First, a unitary transform matrix designed through the LQ decomposition of the MU-MIMO channel is proposed for suppressing MU interference and obtaining the low-triangular MU-MIMO channel effect for each receiving user. A sub channel matrix with a nontriangular channel causes interference from other users to degrade the reception performance for each user. To overcome this problem, we propose a novel cancellation matrix for suppressing nontriangular MU-MIMO channel and composite noise effects. The proposed LQ-based precoder scheme, which is coupled with unitary transformation and cancellation matrices, can be extended for use with multiple users. The LQ-based GSIC precoder for two users can be realized and extended to three or more users by adopting a cascaded structure. The closed-form solution of the proposed LQ-based GSIC precoder is derived using a constraint-based optimization algorithm. Furthermore, suboptimum solutions show that the performance of the proposed GSIC precoder is almost equal to that of the optimum solutions, and that the precoder also has low complexity. Simulation results show that the proposed LQ-based GSIC precoder outperforms conventional precoders and exhibits reliable and excellent sum-rate performance.