Despite its reduced complexity, lattice reduction-aided decoding exhibits awidening gap to maximum-likelihood (ML) performance as the dimension increases.To improve its performance, this paper presents randomized lattice decodingbased on Klein's sampling technique, which is a randomized version of Babai'snearest plane algorithm (i.e., successive interference cancelation (SIC)). Tofind the closest lattice point, Klein's algorithm is used to sample somelattice points and the closest among those samples is chosen. Lattice reductionincreases the probability of finding the closest lattice point, and only needsto be run once during pre-processing. Further, the sampling can operate veryefficiently in parallel. The technical contribution of this paper is two-fold:we analyze and optimize the decoding radius of sampling decoding resulting inbetter error performance than Klein's original algorithm, and propose a veryefficient implementation of random rounding. Of particular interest is that afixed gain in the decoding radius compared to Babai's decoding can be achievedat polynomial complexity. The proposed decoder is useful for moderatedimensions where sphere decoding becomes computationally intensive, whilelattice reduction-aided decoding starts to suffer considerable loss. Simulationresults demonstrate near-ML performance is achieved by a moderate number ofsamples, even if the dimension is as high as 32.
展开▼
