Two novel reordering methods for MIMO sphere detection based on MMSE detection

Sphere detection (SD) can significantly reduce the complexity of multiple-input multiple-output (MIMO) maximum-likelihood (ML) detection, but the complexity of SD is still too high to be applied to the practical system. In order to reduce the complexity of SD, two novel reordering methods based minimum mean square error (MMSE) detection are presented in this paper. Rule-R is based on the reliability of the MMSE detector's decision variable which depends on the magnitude and the SINR of the decision variable. The other method, Rule-M, is a simplified version of Rule-R and only depends on the magnitude of the decision variable. We show that both of them can significantly reduce the complexity with few additional computations. For original SD, both the rules cause no performance degradation. While if some other suboptimal SD were used, the proposed algorithms can both provide computational efficiency and performance improvement.