Iterative soft-decision reed-solomon decoding on partial response channels

Abstract

Since the discovery of turbo codes, iterative decoding has gained enormous momentum. The idea of repeatedly passing information between components of a receiver or decoder to increase the overall system performance has attracted much research effort. In this work, we present an iterative soft-decision decoding architecture for Reed-Solomon (RS) codes on a partial-response (PR) channel. The architecture incorporates a symbol-based a-posteriori probability (APP) detector for the channel, and an enhanced soft-decision RS decoder based upon the recently introduced Koetter-Vardy (KY) algorithm. From the list of candidate RS codewords generated by the KV decoder, we calculate output symbol reliabilities that can be fed back to the APP detector as extrinsic information to be used in a subsequent decoding iteration. Through simulations, we show the efficacy of this approach, especially when the initial KV list size is large. We will also propose ways to modify the decoding scheme in order to beneficially increase the size of the candidate codeword list and thereby improve the overall system performance.