Low Complexity Maximum-Likelihood Based QRD-M for MIMO Systems with S-QAM or APSK

QR decomposition combined with M-algorithm (QRD-M) has received great attention recently due to its low-complexity and near-optimum signal detection performance for multiple-input and multiple-output (MIMO) systems. In this paper, we propose a rotation-based transformation method and apply it to the QRD-M detector for MIMO systems with the commonly used square quadrature amplitude modulation (S-QAM) or amplitude and phase-shift keying (APSK). The QRD-M detector using proposed rotation-based transformation method achieves 50% computation reduction for 4-QAM based MIMO systems compared with the original QRD-M algorithm (25% for those with 16-QAM) without sacrificing any performance. The proposed rotated QRD-M method achieves performance close to that of the maximum-likelihood (ML) detector for 4-QAM modulated MIMO systems with computational complexity lower than the linear minimum mean square error (LMMSE) based detector. The underlying principle of the proposed rotation-based transformation method can be extended to S-QAM (or APSK) modulated MIMO systems employing sphere decoding (SD), list sphere decoding (LSD), list sequential sphere decoder (LISS) (or other quasi-maximum-likelihood detectors), and the full ML based methods in a straightforward manner.