【作者】 张琴；

【导师】 黄胜；

【作者基本信息】 重庆邮电大学 ， 电子与通信工程（专业学位）， 2022， 硕士

【摘要】 极化码是第五代移动通信中用于控制信道的编译码方案,具有巨大的研究价值。极化码的循环冗余校验辅助的串行抵消列表（Cyclic redundancy check Aided Successive Cancellation List,CA-SCL）译码算法的列表大小与纠错性能成正相关,而与复杂度成负相关的关系。本文为了提高极化码的纠错性能,提出了基于耦合的极化码编译码算法。首先设计了一种新的编码方案,每帧信息比特部分根据信道可靠性划分为三部分:高可靠信息比特位集合、较高可靠信息比特位集合和一般可靠信息比特位集合,且较高可靠和一般可靠信息比特位集合的比特位数相等。然后将多个编码帧的较高可靠信息比特位和一般可靠信息比特位进行耦合。接收端对每帧采用首次CA-SCL译码算法,将译码结果满足CRC校验的帧视为译码成功,否则视为译码失败。若全部帧均译码成功就退出译码,否则利用帧间相关性对非连续帧译码失败、连续两帧依次译码失败、连续三帧或三帧以上依次译码失败的三种情况采用帧间比特替换与比特比较相结合的方法,利用译码成功的帧辅助译码失败的帧进行纠正译码,从而提高极化码的纠错性能。仿真结果表明,所提算法与传统的CA-SCL译码算法相比,当误块率为10^-4,码长为1024时有0.55 d B左右的性能增益,与帧间辅助串行抵消列表译码算法相比,约有0.3 d B的性能增益。且当每帧动态冻结位数为80,码长为1024,信噪比为1.5 d B,列表大小为8时,该算法在前两种首次译码情况下的纠错率能达99%,即表明所提算法能够对失败帧进行很好的纠错,使纠错性能得到很大的提高。针对基于耦合的极化码编译码算法在列表数大时复杂度高的问题,本文通过对动态冻结位的首次译码和再次译码的路径度量值和对数似然值进行分析,得出两次译码的数据具有一定的规律和高度相关性,提出了基于耦合的极化码CA-SCL剪枝译码算法。该算法利用帧间耦合性将两次译码的数据相结合来设置再次译码过程中动态冻结位在译码时对错误的、冗余的路径进行剪枝的准则,以降低译码复杂度和提高纠错性能。仿真表明该算法在纠错性能略有提高的同时降低了复杂度,且当列表数越大,该算法降低的复杂度越明显。更多还原

【Abstract】 Polar codes,as a coding and decoding scheme in the control channel of the 5th generation mobile communication standard,which has great research value.The list size of the cyclic redundancy check aided successive cancellation list（CA-SCL）decoding algorithm is positively correlated with error correction performance,but negatively correlated with complexity.To improve the error correction performance of polar codes,the encoding and decoding algorithm of polar codes based on the coupling is proposed in this thesis.Firstly,a new coding scheme is designed,in which the information bits of each frame are divided into three parts according to channel reliability:high reliable bits set,higher reliable bits set,and general reliable bits set,and the higher reliable bits set and generally reliable bits set have the same number of bits.Then,the higher reliable bits set and the generally reliable bits set of multiple encoding frames are coupled.The receiver uses the CA-SCL decoding algorithm to decode each frame the first time.The frame that passes the CRC is regarded as decoded correctly;otherwise,the frame is regarded as a failure.If all the frames are successfully decoded,the decoding will quit;otherwise,inter-frame correlation will be used to combine the method of inter-frame bits replacement and bits comparison for the the three cases of decoding failure of discontinuous frames,sequential decoding failure of two consecutive frames,sequential decoding failure of three or more consecutive frames.The error correction performance of polar codes is improved by using the successfully decoded frame to assist the failed frame to correct the decoded frame.Simulation results show that compared with the traditional CA-SCL decoding algorithm,the proposed algorithm can achieve performance gains of about 0.55 d B when the block error rate is 10^-4 and the code length is 1024.Achieveing performance gains of about 0.3 d B when compared with the inter-frame assisted successive cancellation list decoding algorithm.And when the number of dynamic frozen bits per frame is 80,the code length is 1024,the signal-noise ratios is1.5 d B,and the list size is 8,the error correction rate of the algorithm can reach 99%in the first two cases of the first decoding,which indicates that the proposed algorithm can correct the failed frame well,and the error correction performance is greatly improved.Aiming at the problem of the high complexity of encoding and decoding algorithm of polar codes based on coupling when the list size is large,this thesis analyzes the path metrics and log-likelihood ratio of the first and second decoding of the dynamic frozen bits,it is concluded that the data of two decodings have certain rules and high correlation and proposes a CA-SCL pruning decoding algorithm based on coupling.In order to reduce the complexity of decoding and improve the performance of error correction,the algorithm combines the data decoded twice with the inter-frame coupling to set the criterion of pruning the wrong and redundant paths of dynamic frozen bits in the process of decoding again.Simulation results show that the algorithm reduces the complexity while improving the bits error correction performance slightly,and the algorithm reduces the complexity more obviously when the number of lists increases.更多还原