A New Reduced-Complexity Conditional-Mean Based MIMO Signal Detection Using Symbol Distribution Approximation Technique

In this correspondence, we propose a new conditional-mean based approach for a reduced-complexity suboptimal multiple-input multiple-output (MIMO) detector. In general, the optimal (error-minimizing) metric should take all possible symbol candidates into account and thus exhaustive computations are required. On the other hand, our new approach makes use of approximating the distributions of distinct symbol candidates by a continuous random variable such that the exhaustive summation is replaced by an integration that can be reduced to a simple closed form. The resulting metric depends only on a subset of symbol candidates and thus the overall complexity is reduced considerably. It is found that uniform ring approximation in combination with phase shift keying (PSK) and amplitude-modulated phase shift keying (APSK) achieves a performance close to that of maximum likelihood detection (MLD), while its complexity is a linear order of the modulation multiplicity when the number of transmit antennas is two.