Complexity Reduction Method with Joint Candidate Symbols and Layer Sorting for the Sphere Decoder

The sphere decoder finds the maximum-likelihood (ML) solution with moderate complexity. The computational complexity of the sphere decoder can be further reduced by conventional decoding order scheme sorting the candidate symbols, where the candidate symbols are sorted in ascending order of the distance from the soft-output signal points of a zero-forcing (ZF) receiver. The soft-output signal point of ZF, however, may not be a reliable reference of sorting for finding ML solutions. Moreover, conventional sorting scheme does not consider the searching of the following layer. In this paper, a new sorting scheme is proposed which gives more reliable sorting criteria. It is to sort candidate symbols by considering the number of candidate symbols of the following layer, which are called child symbols. The proposed sorting scheme is combined with layer sorting in order to allow for more reliability at the first detected layer. It is shown that the proposed candidate symbol sorting combined with layer sorting provides around 40% complexity reduction compared with the conventional sorting scheme.