Reliability-Based Soft-Decision Decoding with Iterative Information Set Reduction

In this paper, the reliability-based decoding approach using the reprocessing of the most reliable information set only is extended into the reprocessing of several information sets. At the end of each information set reprocessing, some information bits are delivered by the decoder. Consequently, information sets with decreasing cardinality are considered at each reprocessing stage. A tight upper bound on the error performance achieved by this method is derived. Compared with previously proposed competitive approaches, this new method not only reduces the number of candidate codewords needed to achieve practically optimum decoding, but also allows to achieve nearly any desired level of error performance.