Modified Successive Cancellation Flip Decoder for Polar Codes Based on Gaussian Approximation

Abstract

Compared with the conventional successive cancellation (SC) decoding algorithm, the list decoder effectively improves the block error rate (BLER) performance, but also greatly increases the space complexity. Thus, some scholars have proposed a successive cancellation flip (SCF) decoder that can maintain the similar spatial complexity of the SC decoder while improving the BLER performance of SC decoder. However, the computational complexity of the SCF decoder at low SNR is greater than that of the SC decoder. To overcome the problem and further decrease the complexity, we propose a new decoder called modified successive cancellation flip (MSCF) decoder, which reasonably avoids the unnecessary flip operation of conventional SCF decoder by utilizing a threshold based on Gaussian approximation. Simulation results have shown that the new decoder can significantly reduce the complexity of SCF algorithm with not obvious BLER performance decrease.