一种极化码的优化临界集SCL-Flip译码算法

SCL-Flip Decoding Algorithm for Polar Codes with an Optimized Critical Set

摘要:

针对极化码的传统临界集包含正确路径可能被消除的比特位置问题,提出一种极化码的优化临界集SCL-Flip译码算法。该算法将临界集作为初始临界集,采用高斯近似原理对极化子信道的可靠度进行估计,然后,通过改变初始临界集的选取规则,优化用于串行抵消列表译码中临界集的不完整性,并根据信道可靠度由低到高排序形成优化临界集,该优化临界集能减少重译码次数并有效提高翻转准确性。仿真结果表明,所提算法有较大的性能增益,且翻转次数明显降低。

Abstract:

To address the issue where the traditional critical set (CS) of polar codes includes bit positions that may result in the elimination of the correct path, we proposed a successive cancellation list (SCL)-Flip (SCLF) decoding algorithm with an optimized critical set (OCS-SCLF). In the proposed algorithm, the initial critical set was taken as the starting point, and the Gaussian approximation principle was applied to estimate the reliability of the polarized subchannels. The selection rules of the initial critical set were then modified to address the incompleteness of the CS in SCL decoding. The OCS was constructed and arranged in ascending order of channel reliability, effectively improving the flipping accuracy and reducing the number of re-decoding attempts. The simulation results showed that the proposed algorithm achieved better performance gains while significantly reducing the number of flips.