Improved QRD-QLD Algorithm for Low Complexity MIMO Decoding

In this paper we present a decoding algorithm for multiple-input multiple-output (MIMO) communications. The algorithm is based on the QRD-QLD algorithm (see, e.g., Radosavljevic et al. 2012), proposes an improved path metric computation, and achieves a low probability of error which is close to the single tree search sphere decoder, at low implementation complexity. The proposed improved path metric computation increases the reliability of combining the results of the partial searches into sequences representing the maximum likelihood (ML) and the counter ML hypotheses in soft input MIMO decoding. The excellent performance of the improved algorithm are demonstrated via numerical simulations.