Improved Belief Propagation by Employing the Positive Effects of the Feedback Messages

In this paper, a modification to belief propagation (BP) decoding algorithm is proposed, which is based on extracting the prior messages of each variable node to help the BP decoding, and is particularly effective for low-density parity-check (LDPC) codes with short cycles, where the existence of cycles makes the original BP algorithm perform suboptimal. The proposed algorithm, referred to as "employing the positive effects of the feedback messages (EPEFM)", extracts the positive effects of the feedback messages and then makes use of them as prior messages to assist the decoding of the BP algorithm. Simulation results confirm the effectiveness of our proposed algorithm, which improves the performance in high signal-to-noise-ratio (SNR) region without loss in low SNR region.