Research On The Efficient BP Decoding Of Polar Codes

Polar codes are the first class of channel coding scheme that has been proven to achieve the symmetric capacity of B-DMC.Since Polar code was proposed,it has received extensive attention from academia and industry.The SCL decoding algorithm and various cascaded Polar code schemes,including CRC-Polar code,PC-Polar code,Hash-Polar code,etc.,have effectively improved the bit error rate performance of Polar code,enabling it to Comparable to advanced channel coding schemes such as LDPC codes and Turbo codes.Thesis first introduces the basic principles of Polar codes,including the encoding process of Polar codes and the acquisition and main construction methods of the generator matrix and common decoding algorithms,and compares the bit error rate performance under different code constructions and decoding algorithms.In order to meet the needs of low-latency and high-reliability communication in modern communication systems,thesis mainly studies the BPL decoding algorithm of Polar codes.First,based on the upper bound of the block error rate of the Polar code,a good performance factor graph selection scheme suitable for the BPL decoding algorithm is proposed;then,for the large number logic and the minimum Euclidean distance path selection scheme of the BPL decoding algorithm,A low-latency optimization scheme is respectively proposed;after that,this paper proposes a BPL decoding algorithm based on post-processing;finally,for several factors that affect the BPL decoding algorithm of Polar codes,this paper compares several factors with simulations.The influence of this path selection scheme on the performance of BPL decoding algorithm and the influence of construction method on the performance of BPL decoding algorithm.In order to further improve the performance of the BPL decoding algorithm,this paper also studies the BPL decoding algorithm of the concatenated Polar code,which mainly includes LDPC-Polar code and CRC-Polar code.In order to cope with the low latency,low power consumption,high reliability and other characteristics of the future codec,thesis provides three kinds of Polar code and LDPC code encoding and decoding fusion schemes.First,using the sparse technique of the factor graph of the Polar code,a decoding fusion scheme based on the BP of the Polar code and the LDPC code is given,but the bit error rate performance of the Polar code in this scheme Even the bit error rate performance of the most primitive BP decoding algorithm of Polar code still has a certain gap,in order to further improve the decoding fusion scheme based on BP Polar code and LDPC code Polar Code decoding performance,the research of BPL decoding algorithm is introduced into the decoding,and a decoding based on BPL Polar code and LDPC code is proposed,and related In the performance simulation,we can see that the performance of the optimized solution has been improved to a certain extent.Afterwards,for both the Polar code and the LDPC code are linear block codes,a decoding fusion scheme based on the OSD of the Polar code and the LDPC code is proposed,because OSD is an approximation The decoding algorithm of ML decoding.Therefore,the decoding fusion scheme based on OSD Polar code and LDPC code can obtain good bit error rate performance,but the complexity of this scheme is relatively high.