Iterative reliability-based decoding of low-density parity checkcodes

摘要

In this paper, reliability based decoding is combined with beliefnpropagation (BP) decoding for low-density parity check (LDPC) codes. Atneach iteration, the soft output values delivered by the BP algorithm arenused as reliability values to perform reduced complexity soft decisionndecoding of the code considered. This approach allows to bridge thenerror performance gap between belief propagation decoding which remainsnsuboptimum, and maximum likelihood decoding which is too complex to benimplemented for the codes considered. Trade-offs between decodingncomplexity and error performance are also investigated. In particular, anstopping criterion which reduces the average number of iterations at thenexpense of very little performance degradation is proposed for thisncombined decoding approach. Simulation results for several Gallagern(1963, 1968) LDPC codes and different set cyclic codes of hundreds ofninformation bits are given and elaborated