Research On Parameter Blind Identification And Decoding Technology Of F-LDPC

Channel coding is an important technology to ensure the reliability of data transmission in communication systems.Blind identification of channel coding is an unusual method for a third party to extract information from communication systems in a non-cooperative scenario.This thesis focused on the research of F-LDPC codes,aiming to solve the problems of blind parameters identification and decoding of F-LDPC codes.The main work included the research of encoding and decoding technology,blind identification of parity-check matrix and blind identification of parameters.(1)In the field of F-LDPC encoding and decoding,F-LDPC code could be interpretated as a systematic with serially concatenated parity-check code which has sparse parity-check matrix.F-LDPC encoder is consisted of outer encoder,random interleaver and inner parity generator,and its code rate is highly flexible.Different from classical LDPC codes such as QC-LDPC codes,F-LDPC parity-check matrix is not easy to be stored in a structured way because of the random interleaver.This thesis summarized the association between the structure of parity-check matrix and the parameters based on F-LDPC encoding principle and the generation mechanism of parity-check matrix.Then a real-time parity-check vectors update decoding strategy based on message passing algorithm was introduced,only the parameters and random interleaving vector need to be stored for decoding by this strategy instead of the entire parity-check matrix.(2)In the field of blind identification method of F-LDPC parity-check matrix,the method involves two parts: search of parity-check matrix and restoration of parity-check matrix.The problem of F-LDPC parity-check matrix searching could be transformed into the problem of finding codewords of small weight,in this way the algorithm based on information set decoding could be implemented to search F-LDPC parity-check vectors.Due to the low efficiency of the algorithm at the late stage of searching,this thesis proposed a parity-check matrix restoration algorithm as a supplement to information set decoding algorithm.The combination of two algorithms leads to the complete identification of F-LDPC parity-check matrices for multiple code rates,large code lengths with 10e-4 order of magnitude BER,and small code lengths with 10e-3 order of magnitude BER.(3)In the field of optimization about blind identification method of F-LDPC paritycheck matrix,although the information set decoding algorithm owned certain BER performance,the searching process would be seriously hindered or even stopped completely at large code lengths and higher BERs.In order to improve the efficiency and BER resistance of the blind identification algorithm of F-LDPC parity-check matrix,this thesis innovatively proposed a hamming weight compression method for F-LDPC paritycheck matrix,which optimized the time cost by more than 90% under the same conditions,and achieved the complete identification of F-LDPC parity-check matrix with multiple code rates,large code lengths of 10e-3 magnitude BER and small code lengths of 10e-2magnitude BER.This method provides a feasible solution for the engineering application about blind identification of F-LDPC codes.(4)In the field of blind identification of F-LDPC parameters,the blind identification of F-LDPC parameters includes the repeating parameters,puncturing parameters,and random interleaving.Based on the successful identification of F-LDPC parity-check matrix,this thesis proposed a blind identification algorithm of parameters based on the weight of parity-check matrix which enables the identification of repeating parameter and puncturing parameter.Through the research of F-LDPC code equivalent interleaving,this thesis proposed a blind identification algorithm of equivalent random interleaving based on the structure of parity-check matrix,and verified the equivalence between the identified interleaving and the actual interleaving by theoretical and experimental analysis.