An Improved Successive Cancellation List Flip Decoder for Polar Codes Based on Key Sets
Jingyun Bao, Shunjie Lin, Xingcheng Liu · 2021
Existing studies have documented that flipping operation can improve the performance of polar codes based on CRC-aided successive cancellation list (CA-SCL) decoding algorithm. However, the introduction of the flips greatly increases the complexity and delay of the decoder as well. In this paper, we analyzed the error distribution in the CA-SCL decoding and proposed a new efficient flipping set, key set, to reduce the complexity of the SCL-Flip decoding by narrowing down the size of the flipping set applicable. Different from the critical set which focuses on rate-1 nodes only, the key set can be constructed dynamically based on the results of the original CA-SCL decoding. In addition, we proposed an order-1 SCL-Flip decoder based on the key set to reduce the decoding complexity. Simulation results showed that the key set has a high accuracy of locating the first error and our proposed algorithm has almost no performance degradation compared with the existing SCL-Flip algorithms while the average complexity is reduced significantly.