One-Bit LDPC Message Passing Decoding Algorithm Based on Maximization of Mutual Information

摘要

We propose an discrete message passing decoding algorithm for Low-Density Parity-Check(LDPC) codes under Additive White Gaussian Noise (AWGN) channel assumption. The message-passing decoding map is derived by using an optimum quantizer that maximizes mutual information, using the technique of density evolution. We focus on the case of 1-bit messages The new algorithm has comparable error rate compared with the lowest error rate given by the extensively studied group of Weighted Bit-Flipping(WBP) algorithms. In addition, the new algorithm has very low complexity.