Soft-Input, Iterative, Reed-Solomon Decoding using Redundant Parity-Check Equations

In this work we present a practical approach to the iterative decoding of Reed-Solomon (RS) codes. The presented methodology utilizes an architecture in which the output produced by steps of Belief-Propagation (BP) is successively applied to a legacy decoding algorithm. Due to the highly suboptimal performance of BP conducted on the inherently dense RS parity-check matrix, a method is first provided for the construction of reduced-density, binary, parity-check equations. Iterative decoding is then conducted utilizing a subset of a redundant set of parity-check equations to minimize the number of connections into the least-reliable bits. Simulation results show that performance comparable to (and exceeding) the best known practical RS decoding techniques is achievable with the presented methodology. The complexity of the proposed algorithm is orders of magnitude lower than these existing procedures and permits a practical implementation in hardware.