Noise-Aided Belief Propagation List Bit-Flip Decoder for Polar Codes

Though BP-based polar decoder has low latency, the BLER performance of it does not outperform the CA-SCL decoder with a large list size, which prevents BP-based decoder from its practical implementation. In this paper, we propose a noise-aided belief propagation list bit-flip (NA-BPLF) decoder by merging BPL with artificial noise and bit flipping. In NA-BPLF, when CA-BPL decoding fails, the error-prone bit is chosen for the bit flipping operations based on the decoding results generated by the previous CA-BPL decoding process, and then artificial noise is superimposed. Simulation results show that NA-BPLF achieves better BLER performance which is close to large list size CA-SCL decoder while still keeping the low latency.