A new faster sphere decoder for MIMO systems

Sphere decoding technique is an efficient fast algorithm for maximum likelihood (ML) detection. In this paper, we propose a new faster sphere decoding algorithm for multiple-input multiple-output (MIMO) systems that its complexity is much less than the conventional algorithm. In. our method, the initial selection of sphere radius is not important. It will be shown that both the performance and the complexity of the proposed algorithm are simultaneously controlled by two parameters. Therefore, there is a trade-off between the performance and the complexity in the proposed algorithm. By computer simulations, we show that the performance of our method is very close to ML decoding algorithm. Furthermore, if we tolerate some degradation in the performance, we can speed up the proposed algorithm to ten times. Our results show that even in the fastest case, the novel decoder outperforms the V-BLAST decoding algorithm.