Research On Construction Of QC-LDPC Codes Based On Genetic Optimization

With the rapid development of wireless communication technology,the public has higher requirements for communication speed and communication quality of information transmission.At present,the 5th gen-eration(5G)mobile communication has begun to be commercialized on a large scale in my country,and the application scenarios are constantly enriched.5G services are not only applied to smartphones,but also serve a variety of 5G devices,such as extended reality,cloud gaming devices,in-vehicle devices,uncrewed aerial vehicles,etc.5G advanced(5G+)and the 6th generation(6G)mobile communication put forward new requirements for information transmission.Higher transmission rate,higher reliability,and lower delay become the development direction of channel coding,which also puts forward higher requirements for the performance of low-density parity-check(LDPC)codes in the 5G service channel coding scheme.This paper studies the coding optimization and construction of quasi-cyclic LDPC(QC-LDPC)codes.A comprehensive analysis of QC-LDPC codes in binary and non-binary fields is carried out,and genetic algorithm is introduced to optimize the coding structure,the main work and achievements are as follows:The paper first introduces the research background.The development of wireless communication tech-nology,LDPC decoding method,construction method and performance analysis method are briefly intro-duced.Then,the definition,description method,basic construction method and coding and decoding al-gorithms of binary LDPC codes and non-binary LDPC codes are described in detail,and the check matrix structure,code length and code rate compatible method,and specific coding implementation of 5G-LDPC codes are introduced.These provide a theoretical basis for the following research.Then,the paper focuses on the optimization scheme of QC-LDPC codes based on genetic algorithm.The paper introduces the basic steps of genetic algorithm,and designs a scheme of applying genetic algorithm to QC-LDPC code optimization.The key steps in the algorithm,such as population initialization,fitness function calculation,selection,mutation and crossover,are described pertinently for the construction of QC-LDPC codes.Then,the high-speed decoding simulation test platform of binary LDPC code is introduced,which uses the single instruction multiple data(SIMD)instruction set and multi-core processor to accelerate the parallel test of decoding simulation.In this paper,the appropriate code sets are selected in the BG1 and BG2 matrices of 5G-LDPC codes respectively for the simulation experiments and performance comparisons of code length compatibility and code rate compatibility.The simulation results show that the codes optimized by genetic evolution can obtain significant performance improvement,and the performance gain at the design target of the optimization algorithm can reach up to 0.5d B.Next,the paper further extends the genetic optimization scheme of QC-LDPC codes to the design of non-binary LDPC codes.Firstly,the coding construction processes of non-binary LDPC codes and non-binary QC-LDPC codes are analyzed,and the construction method of non-binary LDPC codes is given.Then the specific implementation schemes of genetic algorithm in optimizing the replacement matrix and non-binary check matrix of LDPC codes are described in detail,and the genetic optimization method of binary QC-LDPC code is improved considering the replacement matrix,basis matrix and both of these in non-binary QC-LDPC codes.In this paper,some representative non-binary LDPC codes are selected for construction optimization and simulation tests.The paper also proposes a code length extension method,and extends the multi-code length for some non-binary LDPC codes.The simulation results show that the optimized algorithm can effectively improve the decoding performance,and the non-binary LDPC codes after code length extension can still obtain significant decoding gain.In addition,the paper presents a construction design scheme of a non-binary QC-LDPC code,and uses this scheme to construct a non-binary QC-LDPC code whose performance is close to the existing BDS protocol in GF(64).Finally,the paper also discusses the genetic algorithm coding optimization of QC-LDPC codes com-bined with high-order modulation.Firstly,the high-order coding and modulation method of LDPC codes is introduced,and the bit interleaved coded modulation(BICM)system and the bit interleaving scheme of5G-BICM are described.Furthermore,the high-order coding modulation scheme based on non-binary LDPC coding is studied.The paper selects the codes based on BG1 matrix of the 5G-LDPC codes for single coding optimization,multi-code length compatibility optimization and multi-code rate compatibility optimization.And some non-binary QC-LDPC codes are selected to optimize the structure of the coding modulation sys-tem.The simulation results show that the genetic algorithm can effectively optimize the construction design of QC-LDPC codes and non-binary QC-LDPC codes combining high-order modulation,so as to achieve the purpose of improving the overall performance of the coding and modulation system.