Weighted bit-flipping decoding for product LDPC codes

Abstract

In this work, we propose a weighted bit-flipping (WBF) decoding for product LDPC codes of which decoding complexity is lower than the belief-propagation (BP) decoding. Two distinct types based on page computations and row/column computations are proposed. In addition, we address an issue of hard decision algorithm for product LDPC codes suffering a performance degradation at high iteration numbers. We introduce a threshold of flipped bits to adjust the performance at high iterations. Although the performances of both proposed decoding are worse than that of the BP algorithm, it greatly reduces the decoding complexity.