Reduced complexity list sphere decoding for MIMO systems

Abstract

Depth-first sphere decoding of MIMO systems has near maximum likelihood performance with reasonable computational complexity. In this paper, lower complexity depth-first sphere decoding and list sphere decoding algorithms are proposed. Several criteria for re-ordering the search dimensions are proposed. The proposed sphere decoders are shown to have a significantly reduced decoding complexity at low SNRs. To further reduce the complexity at high SNRs, the point search-space at each ordered dimension is adaptively reduced. Further reductions in the decoding complexity are achieved by inter-layer interference cancellation. It is shown that the proposed sphere decoding algorithms maintain their near-optimal performance, concurrently with a significant complexity reduction, over a wide SNR range.