A Simplified Unbiased MMSE Metric Based QRDM Decoder for MIMO Systems

In contrast to the conventional QR-decomposition M (QRDM) algorithm, the unbiased MMSE metric based QRDM algorithm exploits MMSE metric instead of the maximum likelihood (ML) metric, leading to enhanced performance with less complexity. However, for an efficient hardware implementation, the sorting stages involved at every QRDM detection stage can still be quite time and resource consuming. Hence, we propose a simplified QRDM approach, denoted as S-QRDM, which only performs the full enumeration (FE) on certain number of top search layers, while utilizing the single enumeration (SE) for the remaining search layers. Therefore, the need for the sorting process for those layers is eliminated and the required number of multipliers is reduced simultaneously. The proposed S-QRDM detector has been implemented on a state-of-the-art Xilinx FPGA device and the results indicate that the proposed approach is more efficient in either logic or computational FPGA resources compared to the original algorithm.