Parallel QRD-M Encoder for Decentralized Multi-User MIMO Systems

In this paper, we propose a parallel QRDM encoder (PQRDME) for multi-user MIMO (MU-MIMO) systems. The proposed algorithm transforms the full tree-search problem of the conventional QRDME algorithm into parallel partial trees that are processed in parallel, leading to a tremendous increase in the encoding throughput. The proposed algorithm outperforms the fixed-complexity sphere encoder (FSE) and performs close to the optimum performance for several scenarios. For instance, in a 4×4 MUM-MIMO system, the proposed PQRDME lags the optimum performance by 0.3 dB while outperforming the FSE by 2.3dB at a target BER of 10^-4. In this case, the proposed algorithm also doubles the encoding throughput of the conventional QRDME algorithm.