A Generalised Low-Complexity SISO Sphere Decoder for Lattice-based Communication Systems

Abstract

Soft-input soft-output sphere decoders are known to achieve near-maximum likelihood performance with a reasonable complexity, in lattice-based communication systems. They are designed by combining a list sphere decoder with the a posteriori probability technique. In this paper, we first derive generic linear equations that bind a lattice point to a received codeword in single or multiple antennae lattice-based communications. We then propose a novel list-sphere decoding algorithm that is generalised for any kind of two-dimensional constellations, any type of channel matrices, and any number of transmit or receive antennae. This algorithm generates an optimal list of possible transmit codewords, which are the closest to the one received in terms of Euclidean distance, with a computational complexity independent of the initial sphere radius. Results show that our algorithm exhibits a lower complexity than the previously-designed algorithms.

Document Downloads