Improved QRM-ML detection with candidate selection for MIMO multiplexing systems

This paper proposes an improved QR decomposition associated M algorithm for maximum-likelihood detection (QRM-ML) with candidate selection in multiple input multiple output (MIMO) multiplexing systems. In the proposed algorithm, only the points falling into a sub-set of constellation are selected as the candidates for each transmitted signal and included in the squared Euclidean distance calculation. Unlike the existing algorithms where the sub-set is fixed based on quadrant, the sub-set here is determined by the circle with center at the signal estimate and a pre-determined radius. By computer simulations, it is shown that the proposed QRM-ML algorithm can achieve the performance close to that of the existing QRM-ML algorithms with significantly reduced complexity.