Authors: Hannes Bartz Vladimir Sidorenko
Publish Date: 2016/03/23
Volume: 82, Issue: 1-2, Pages: 449-467
Abstract
An efficient interpolationbased decoding algorithm for hfolded Gabidulin codes is presented that can correct rank errors beyond half the minimum rank distance for any code rate 0le Rle 1 The algorithm serves as a list decoder or as a probabilistic unique decoder and improves upon existing schemes especially for high code rates A probabilistic unique decoder with adjustable decoding radius is presented The decoder outputs a unique solution with high probability and requires at most mathcal Os2n2 operations in mathbb F qm where 1le sle h is a decoding parameter and nle m is the length of the unfolded code over mathbb F qm An upper bound on the average list size of folded Gabidulin codes and on the decoding failure probability of the decoder is given Applying the ideas to a list decoding algorithm by Mahdavifar and Vardy Listdecoding of subspace codes and rankmetric codes up to Singleton bound ISIT 2012 improves the performance when used as probabilistic unique decoder and gives an upper bound on the failure probabilityThe authors would like to thank Gerhard Kramer Joschi Brauchle and Johan S R Nielsen for fruitful discussions and helpful comments H Bartz was supported by the German Ministry of Education and Research in the framework of an Alexander von HumboldtProfessorship
Keywords: