Study On The Algebraic Construction And Decoding Algorithm Of LDPC Code

Low density parity check(LDPC)code is a kind of linear block error correction code,which has been widely studied because of its close to Shannon channel capacity.In this paper,the construction and coding of LDPC code check matrix,as well as the decoding algorithms of BP and IDs are studied and improvedFirstly,the improved QC-LDPC construction and coding method based on Fibonacci sequence is studied.On the basis of introducing the construction method and basic coding method of LDPC check matrix,the construction method and coding method of QC-LDPC based on Fibonacci sequence are put forward.The check matrix constructed by this construction method does not contain four rings,and saves storage space without loss of performanceSecondly,the improvement of LDPC decoding algorithm based on normalized minimum is studied.Based on the comparison of llr-bp and min sum decoding algorithms,an improved NMS decoding algorithm is proposed for the reason of performance loss of Min sum algorithm.Based on the normalized min sum algorithm,the improved decoding algorithm uses the density evolution theory to obtain better decoding performance by dynamically selecting the normalization factor.Thirdly,the improvement of LDPC decoding algorithm based on dynamic scheduling strategy is studied.Based on the analysis of two classical LDPC dynamic scheduling decoding algorithms,RBP and nw-rbp,an improved RBP algorithm is proposed to improve the high BER platform caused by "greedy" of RBP algorithm,which can effectively prevent the high BER platform caused by the greedy group.Fourthly,the LBP decoding algorithm based on reliability is studied.Based on the introduction of the classical LBP decoding algorithm,a reliability based LBP decoding algorithm is proposed.It measures the reliability of variable nodes by the log likelihood ratio information obtained from the channel.Each variable node can estimate the number of decoding iterations by its own reliability.Compared with the traditional flooding scheduling algorithm and LBP algorithm,the algorithm has faster convergence speed,but slightly increased complexity;compared with vc-rbp algorithm,in the same decoding performance,it greatly reduces the computational complexity.