A tree-based ML decoding algorithm with adaptive thresholding |
In this work, we propose a maximum-likelihood (ML) soft-decision decoding algorithm of linear block codes based on an enhanced tree-based search algorithm. The algorithm considers each codeword as a tree branch and explores the subtree corresponding to the more reliable basis (MRB) of received vectors. By making use of the pre-defined ensemble branches and their costs, the branch evaluation process is effectively controlled by adaptive thresholding. This paper details the implementation of the proposed algorithm, compares the performance with the classic ordered statistic decoding (OSD) algorithm, and discusses the general decoding complexity and search effort in comparison to that of the OSD algorithm.
Cai J, Tomlinson M, Tjhai CJ, Yang L