一种极化码的优化临界集SCL-Flip译码算法

A SCL-Flip Decoding Algorithm of Polar Codes with the Optimized Critical Set

摘要:

针对极化码的传统临界集(Critical Set, CS)包含正确路径可能被消除的比特位置的问题,提出一种极化码的优化临界集SCL-Flip (Successive Cancellation List Bit-Flip with the Optimized Critical Set, OCS-SCLF)译码算法。该算法将CS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(Successive Cancellation List, SCL)译码中CS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(Optimized Critical Set, OCS),该OCS能减少重译码次数并有效提高翻转准确性。仿真表明:所提算法有较大的性能增益,且翻转次数明显降低。

Abstract:

In order to improve the problems that the traditional critical set (CS) of polar codes contains the bit positions where the correct path may be eliminated, an SCL-Flip with the optimized critical set (OCS-SCLF) decoding of polar codes is proposed. In the proposed algorithm, the critical set is taken as the initial critical set, the Gaussian approximation principle is used to estimate the reliability of the polarized subchannel, then the incompleteness of the CS in the SCL decoding is optimized by changing the selection rules of the initial critical set,, and the optimized critical set(OCS) is constructed and arranged according to the ascending order of the channel reliability, the OCS can effectively improve the flipping accuracy and reduces the number of redecoding. The simulation results show that the proposed algorithm has a better gain of the performance, furthermore, the number of the flip is significantly reduced.