Complex sphere decoding with a modified tree pruning and successive interference cancellation

A novel complex sphere decoder is proposed in this paper, which achieves a lower complexity than existing techniques with the aid of successive interference cancellation (SIC), statistical pruning strategy (SPS) and a modified probabilistic tree pruning (MPTP) in the search strategy. In the proposed method, enumeration schemes can be avoided once the best partial path metrics (PPMs) fail to satisfy the constraint obtained by the MPTP for a given layer. In addition, the SIC provides a simple method to obtain the possible best candidate for each layer. This strategy effectively eliminates the extra branches without enumeration, if no candidates are available for each layer. Simulations illustrate that the proposed algorithm can reach near-maximum likelihood (ML) performance with a reduction in computational complexity.