Sorting operation is one of the main bottlenecks for the successive-cancellation list (SCL) decoding. This paper introduces an improvement to the SCL decoding for polar and pre-transformed polar codes that reduces the number of sorting operations without degrading the code's error-correction performance. In an SCL decoding with an optimum metric function we show that, on average, the correct branch's bit-metric value must be equal to the bit-channel capacity, and on the other hand, the average bit-metric value of a wrong branch can be at most zero. This implies that a wrong path's partial path metric value deviates from the bit-channel capacity's partial summation. For relatively reliable bit-channels, the bit metric for a wrong branch becomes very large negative number, which enables us to detect and prune such paths. We prove that, for a threshold lower than the bit-channel cutoff rate, the probability of pruning the correct path decreases exponentially by the given threshold. Based on these findings, we presented a pruning technique, and the experimental results demonstrate a substantial decrease in the amount of sorting procedures required for SCL decoding. In the stack algorithm, a similar technique is used to significantly reduce the average number of paths in the stack.
翻译:排序操作是连续取消列表( SCL) 解码的主要瓶颈之一 。 本文将改进 SCL 解码极地和预先转换的极地代码, 从而在不降低代码错误校正性能的情况下减少分类操作的数量。 在具有最佳度量函数的 SCL 解码中, 我们显示, 平均而言, 正确的分支的位数值必须等于比特通道断裂速度, 而另一方面, 错误分支的平均位数值可能最多为零 。 这意味着错误路径的偏差路径部分度量值偏离了比特通道能力的部分加码值。 对于相对可靠的位通道, 错误分支的位数的比特度值会变得非常负值, 从而使我们能够检测和绘制这样的路径。 我们证明, 在比特道断裂速度更低的阈值下, 正确路径的概率比给定阈值指数下降。 基于这些发现, 我们展示了一种修路道技术, 而实验结果显示, 在比特通道的轨迹中, 正在大幅下降的轨算法。