Combined SCF and SCAN Algorithms for Soft-Output Polar Decoding

Abstract

Polar code has lower coding and decoding requirements, linear complexity, and superior channel performance, so it is proven to be close to the Shannon limit for 5G communication. Successive cancellation (SC) is a polar code decoding method, but its performance has a lot of room for optimization, and flip methodology is one of its basic optimizations. Successive Cancellation Flip (SCF) has a good error correction capability. But due to the repetition of the SC decoder during computation, it has a very long latency. Moreover, SCF output is a hard decision, so it is not capable of soft input and soft output decoder. To allow SCF to be used in more decoding environments, the Hybrid SCF/SCAN proposed in this paper can freely select soft output and hard output, and the proposed architecture can reduce the problem of repeatedly computing SC decode.