Linear Block Code Research On Decoding Algorithm Of Polar Code And HARQ Technique Of LDPC Code

Polar code and low density parity check(LDPC)code are two of the most valuable encoding methods of linear block codes in communication field,and have been adopted by 5G standard as the channel coding schemes of control channel and shared channel,respectively.LDPC code is also the most important channel coding scheme in Wi-Fi standard.Polar code is mainly used as the encoding scheme for short and medium code transmission in 5G standard.However,when the code length is short,the channel polarization is incomplete,so that the reliability of polar code decreases significantly.In addition,the commonly used successive cancellation(SC)decoding algorithm,CRC-aided successive cancellation list(CA-SCL)decoding algorithm cannot correct the decoding errors that have occurred during the decoding process.Once a burst error occurs,the whole codeword cannot be successfully decoded,which further limits the block error rate(BLER)performance of polar code.Focus on the typical scenarios in 5G transmission and the characteristics of polar code,this thesis proposed a successive cancellation list flipping algorithm for short polar code based on row weights of generator matrix.In this thesis,the relationship between the row weights of generator matrix and the position distribution of the first error bit is revealed by the characteristics of SCL decoding algorithm and sufficient simulation experiments.Then,this relationship is applied to bit flipping algorithm.By taking advantage of message passing phenomenon,if the decoding result of the first error bit is flipped deliberately during decoding process,and then all of the following decoding results will be affected by it.Therefore,the decoding errors can be corrected with a certain probability.The proposed algorithm is suitable for short polar code and can accurately select the first error bit.Compared with the existing algorithms,the BLER performance can be improved up to 0.3 d B in the case of short code length,and the decoding complexity is slightly lower than the existing algorithms in the case of low signal to noise ratio(SNR)region.The proposed algorithm has great prospect in short packet transmission scenarios such as machine type communication(MTC)and so on.LDPC code,which is a traditional linear block code,has been widely used.The design of LDPC code is mature in 5G standard,the check matrix and HARQ algorithm have been refined.The design of LDPC based HARQ method is a hot topic in debate of the next generation Wi-Fi standard.In this thesis,the connection of bit error probability within a codeword with degree and short loop distribution is analyzed in the beginning.Then,concerned with the feature of check matrix in Wi-Fi standard,the degree and short loop distribution based HARQ algorithm for LDPC code is proposed.Simulation results show that compared with the existing HARQ algorithm,the proposed algorithm can obtain up to 2.5 d B energy gain under the same throughput performance in Gaussian channel.Under Wi-Fi standard channel conditions,compared with the proposed algorithm,the maximum energy gain can be achieved by 0.5 d B under the same throughput performance.Concerning the proposed algorithm has little modification to the existing standard,it is convenient to realize the downward compatibility of standards,which makes it promising to be applied into practical scenarios in the future.