Research on Decoding for Polarization-Adjusted Convolutional Codes with New Node Metric
Jian Qiao, Lizhe Liu, Jinbo Zhang +4 more
In the Shannon Lecture at the 2019 International Symposium on Information Theory (ISIT), Professor Arikan proposed a new coding scheme called polarization-adjusted convolutional (PAC) codes. These codes exhibit excellent error-correction performance for short-to-medium block lengths. Currently, the mainstream decoding algorithms for PAC codes mainly include sequential decoding and list decoding. Fano decoding, a type of sequential decoding, offers good error-correction performance; however, its time complexity is relatively high due to the presence of extensive backtracking processes. In this paper, a fast Fano decoding algorithm based on backtracking optimization mechanism for PAC codes is proposed. The improved algorithm we put forward mitigate the time complexity of the decoding process from two aspects: path backtracking and threshold updating strategy. It enhances the error-correcting performance of the Fano decoding algorithm by introducing the first error node metric and the second error node metric. With the efforts in the above aspects, our decoding algorithm achieves good performance in both complexity and error-correction capability.
IEEE International Conference on Communication Technology (ICCT)•Oct 16, 2025•0 citations