Design And Application Of LDPC Encoding And Decoding Algorithms

Since the end of 1990 s,in the wake of rapid increase of multimedia application re-quirements,the wireless image transmission system has put forward higher requirements on anti-error performance.As an anti-error transmission technology,the joint source chan-nel coding(JSCC)has been rapidly developed.Moreover,as a potential channel coding technology,LDPC code has been widely researched in JSCC systems.On the other hand,the explosive growth of information has put forward higher requirements for the capacity and reliability of data storage systems.However,the performance of mainstream error correction codes such as Reed-Solomon(RS)code and Bose-Chaudhuri-Hocquenghem(BCH)code etc.,have reached or approached their limits.The LDPC code as an im-portant error-correcting coding technology with excellent performance under the iterative decoding is widely studied in various mass storage systems,such as semiconductor stor-age system,high density optical storage system and deoxyribonucleic acid(DNA)data storage system,etc.In view of the important roles of LDPC codes in the wireless com-munication and data storage technologies,this dissertation focuses on the study of LDPC coding and decoding algorithms for JSCC transmission systems and data storage systems.The main contributions of this work include the following aspects:1.LDPC code design for the JSCC transmission systemFirstly,considering the double protograph low-density parity-check(DP-LDPC)code based JSCC system sufffers from high error floor for the source sequence with short block length and high source probability,a joint optimization method of finite length DP-LDPC code is proposed,which can obtain low error floor with satisfied waterfall region perfor-mance.In addition,an adaptive rate allocation strategy based on fuzzy logic control is adopted to further improve the transmission reliability of DP-LDPC short codes.Secondly,considering the spatially coupled LDPC(SC-LDPC)code based JSCC scheme adopts fixed and short coupling length and sliding window decoding strategy,leading to limited improvement of the overall performance,the anytime coding tech-nique is introduced into the SC-LDPC based JSCC system,named as joint source channel anytime coding(JSCAC).In the proposed JSCAC,the exponential distribution coupling method and the joint extended window decoding(JEWD)adopted in JSCAC scheme can ensure that the transmitted but not fully recovered sub-information blocks with high source probability and short length can be corrected quickly.In addition,the proposed partial joint extended window decoding(PJEWD)strategy can further reduce the error propaga-tion between the source decoder and channel decoder,which can also improve the error correction ability of the decoder.2.LDPC code design for the DNA data storage systemIn order to correct the asymmetric substitution error in the DNA data storage,a hy-brid coding system consisting of modified variable-length run-length limited(VL-RLL)code and protograh LDPC code was proposed.The improved VL-RLL code is designed to meet the biological constraint of DNA sequence and obtain the mapping potential close to the limit.In terms of the protograh LDPC code design,an improved protograph extrin-sic information transfer(PEXIT)algorithm for the asymmetric substitution channels was firstly designed,and then series of protograh LDPC codes are proposed accordingly for different sequencing channels.The simulation results show that the optimized protograph LDPC codes have better error performance than the traditional protograh LDPC code and the codes used in the existing DNA data storage systems.3.Application of BP decoding and scheduling strategies on the high-density parity-check(HDPC)codeIn the application of LDPC decoding algorithm,a perturbed adaptive belief propa-gation(P-ABP)algorithm is proposed.The propsed P-ABP is used to improve the soft-in-soft-out(SISO)decoding performance of HDPC codes,which are commonly used in data storage systems.The key idea of traditional ABP is to sparsify certain columns of the parity-check matrix corresponding to the least reliable bits with small log-likelihood ratio(LLR)values.This sparsification strategy may not be optimal when some bits have large LLR magnitudes but wrong signs.Motivated by this observation,a perturbed ABP(P-ABP)algorithm is proposed to incorporate a small number of unstable bits with large LLRs into the sparsification operation of the parity-check matrix.In addition,an improved partial layered scheduling and a hybrid dynamic scheduling strategy are also applied to further enhance the performance of P-ABP.Simulation results show that the proposed de-coding algorithms lead to improved error correction performances and faster convergence rates than the prior-art ABP variants.