Improved Stack Decoding for PAC Codes

Abstract

Several classical decoding algorithms, such as Fano decoding, list decoding and list Viterbi decoding algorithm have been proposed for polarization-adjusted convolutional (PAC) codes by some scholars. Inspired by the decoding algorithms mentioned above, two algorithms called stack Viterbi decoding (SVD) and critical set-aided stack decoding are presented in this paper. Simulation results show that performance of the proposed algorithm with stack depth of 100 slightly outperforms that of the conventional stack decoding algorithm with the same stack depth at the cost of higher complexity, and the sorting complexity of conventional stack decoding (CSD) can be further reduced by means of the proposed critical set-aided stack decoding algorithm.