Research On Construction And Decoding Algorithm With Low Error Floor Of LDPC Codes

In recent years,with the rapid development of the channel coding technology,low density parity check(LDPC)code has been widely concerned by researchers all of the world as an approaching code to Shannon limit.And many successful applications are obtained in the communication system of reliable information transmission.Error floor is an important issue in the theory of LDPC codes and the research of iterative decoding algorithms.It is characterized that the error rate in high SNR region suddenly drops at a rate much slower than that in the region of moderate SNR.In this background,some of the most extensive work on constructing and decoding algorithm with low error floor of LDPC codes are discussed,and the main works is summarized as follows:(1)Based on the construction of LDPC codes,an algorithm to improve the minimum distance of codes by eliminating distance sets is proposed for irregular LDPC code.The idea of distance sets elimination algorithm first to add a small amount of check nodes in Tanner graph corresponding to the code word,then let these check nodes connect with some variable nodes in distance sets.In order to achieve the purpose of eliminating the distance sets and reduce the error level.(2)Based on the decoding of LDPC codes,two new improved decoding algorithms are proposed for the different scheduling strategies of the LDPC codes:The first algorithm is based on variable nodes,Group Belief Propagation(GBP).According to the difference between the messages and the hard decision values,the variable nodes are divided into reliable nodes and unreliable nodes.And the unreliable nodes are priority to update.Further,the effects of threshold for judging the group on GBP strategy is analyzed.Simulation results show that the proposed algorithm can achieve better convergence and decoding performances compared to three classical scheduling algorithms.Meanwhile,It can overcome the error floor phenomenon caused by the trapping set effectively.The second improved algorithm is based on the check node,and the Check Node of Group Belief Propagation(CN-GBP)algorithm is proposed.At first,the check nodes are also divided into two different group depending on reliability,and the unreliable nodes are updated first.The simulation results show that the CN-GBP can be better than Layered BP(LBP)in the convergence rate and the decoding performance.What's more,CN-GBP can overcome the trap sets in a certain degree.