Construction of Shortened Systematic PAC Codes Based on Monte-Carlo Algorithm

Abstract

Polarization transform limits the code length of polarization-adjusted convolutional (PAC) codes to a power of 2, which hinders the flexible use of PAC codes. Shortening is an important method to change the length of the PAC codes. Systematic code makes it possible to select the shortened bit flexibly. In this paper, we propose a flexible code length scheme for shortened systematic PAC codes, where the proposed scheme generates frozen set and shortened set based on Monte-Carlo algorithm. Simulation results show that the Monte-Carlo algorithm can significantly improve the performance in frame error rate (FER), and the proposed shortening scheme can slightly improve the performance in FER.