Research On Encoding And Decoding Of LDPC Code And Its Application In 5G System

Low Density Parity Check(LDPC)code is a linear block code which has sparse parity check matrix and excellent performance.In New Radio(NR)standardization,LDPC code has been adopted as data channel coding scheme for Mobile Broadband Enhancement(eMBB)scenario,which is one of the three scenarios in NR.Therefore,the research on encoding and decoding of LDPC code and its application in 5G system makes sense.The content of this thesis is divided into two parts:LDPC code based on 5G NR and LDPC code belief propagation(BP)decoding algorithm and its improved algorithm.In LDPC code based on 5G NR,several problems in standardization are studied.Firstly,the structure of parity check matrix is crucial to the performance of LDPC code,this thesis studies parity check matrix structure and encoding algorithm of LDPC code.NR standardization uses quasi-cyclic LDPC(QC-LDPC)code whose parity check matrix is characterized by basic matrix and expansion factor,and its basic matrix has special double diagonal structure and single diagonal structure.Based on this feature,this thesis proposes a encoding algorithm that directly uses basic matrix.Secondly,this thesis studies false alarm rate(FAR)performance of LDPC code itself and FAR performance after adding cyclic redundancy check(CRC).Data communications demands integrity and correctness of data transmission,the FAR performance requirements are usually achieved by CRC.But LDPC code is a linear block code which the codeword is constrained by parity check matrix,so it has FAR performance,hence this thesis researches related factors for LDPC code itself FAR performance and FAR performance after adding CRC.Thirdly,this thesis studies interleaver problem in high order modulation of LDPC code.On one hand,the reliability of constellation points at different positions in high order modulation is different.On the other hand,the reliability of variable node at different locations varies greatly caused by the special structure of parity check matrix.Therefore a interleaver which give more reliability for variable nodes that have large degree is proposed,and simulation shows this scheme can improve bit error rate performance.LDPC code BP decoding algorithm and its improved algorithm mainly studies probability domain BP algorithm and logarithm domain BP algorithm(SPA)of LDPC code at first.SPA algorithm’s check node operation is complex,so MSA algorithm and its modification algorithm are introduced,and these algorithms are analyzed through simulations.Next,the layered BP algorithm(LBP)which has the advantages of hierarchical parity check matrix,faster convergence and lower error platform is introduced.Its principle and process are introduced.Its simplified algorithm LMBP and its modified algorithm are also introduced and analyzed through simulations.Finally,the relationship between the number of established parity check(PC)equations and the number of iterations is studied through simulations.Besides,the relationship between output amplitude average change value of variable nodes and the number of iterations is also studied through simulations.For the low signal to noise ratio,this thesis proposes PC_ET algorithm which uses the number of establishment of PC equation as the indication of early termination.This thesis also proposes PC&VN_ET algorithm which combines the establishment of the number of PC equations and output amplitude average change value of variable node as the indication of early termination,finally the performance of PC ET algorithm and PC&VN ET algorithm are analyzed and compared by through simulations.