Research On Joint Source And Channel Decoding Algorithm Based On Polar Codes

Polar codes,which have only been proposed for a little over a decade,have already been applied in the fifth-generation mobile communication（5G）standard to provide solutions for high reliability and high spectral efficiency system requirements.Currently,Polar codes are also one of the important candidate technologies for 6G communication and have a very broad application prospect.In traditional communication systems,the source and channel are mostly designed separately,but Shannon’s theorem shows that joint source-channel design is more effective in approaching the channel capacity limit under finite code length.Although there have been many excellent achievements in the field of joint source-channel design,research on Polar codes is still in its infancy.Therefore,exploring source-channel joint systems based on Polar codes is of great significance for improving system performance.This thesis mainly focuses on the research of the joint source and channel decoding（JSCD）algorithm based on polar codes,and the specific work is summarized as follows:（1）This thesis has studied the coding and decoding theory of the source and channel parts of polar codes and optimized the structure of the JSCD system.By analyzing different construction methods and encoding schemes,the system framework of the joint source-channel decoding in this thesis is determined.Specifically,the Bhattacharyya parameter method is used to construct the source polar codes,and the gaussian approximation（GA）method is used to construct the channel polar codes.The encoding method of systematic polar codes（SPC）is adopted to improve the system’s ability to resist channel noise,and the joint-successive cancellation list（J-SCL）algorithm is used for decoding.（2）To address the potential problem of missed codewords in the JSCD system based on polar codes,a CRC-aided joint-successive cancellation list algorithm（CRC-JSCL）is proposed by utilizing the advantages of low missed detection rate,low complexity,and fast check of CRC codes.This algorithm precisely outputs the decoding codewords and effectively improves the bit error rate（BER）performance of the J-SCL algorithm.Simulation results show that for Polar codes with a source code length of 512,a source compression rate of 0.6,and a list size of 32,introducing a 16-bit CRC for decoding can obtain a gain of approximately 0.22 dB compared with the original joint decoding algorithm under the same conditions when the BER is10^-4.（3）In order to further improve the error correction capability and the utilization of decision bits in the joint decoding algorithm,this thesis proposes a CRC aided joint successive cancellation list flip（CA-JSCLF）decoding algorithm based on bit flipping.This algorithm constructs a flip set-based on the joint decoder by reducing the search range of the traditional flip set.During decoding,the algorithm re-decides the codewords in the flip set to cut off the error propagation caused by bit misjudgment in the CRC-JSCL algorithm,balancing algorithm complexity and error rate performance.Simulation results show that for a polar code with a source code length of 512,source compression rate of 0.6,and list size of 8,with the maximum number of flips set to 32,the proposed CA-JSCLF algorithm can achieve a gain of about 0.15dB compared to the CRC-JSCL algorithm under a BER of 10^-4.