Information-Propagation-Based Scheduling for the Fast Convergence of Shuffled Decoding

Abstract

This letter presents a scheduling scheme for the fast convergence of shuffled decoding. The messages from the variable node group that is expected to propagate the most reliable information are updated first. Mutual information is used to measure the information reliability. Simulation results show that the average number of iterations is reduced by adopting the proposed schedule, compared with previous scheduling schemes. In addition, information-theoretic analysis results show that mutual information in the proposed scheme increases faster than in the previous scheme.