Self-Corrected Min-Sum decoding of LDPC codes

TL;DR

This paper introduces a simple self-correction method for Min-Sum decoding of LDPC codes that enhances performance by erasing unreliable messages, achieving near Sum-Product decoding accuracy with low complexity.

Contribution

The proposed method modifies variable node processing in Min-Sum decoding without correcting check node approximations, leading to improved decoding performance.

Findings

Performs close to Sum-Product decoding in simulations

Maintains low complexity of Min-Sum decoding

Robust against noise variance estimation errors

Abstract

In this paper we propose a very simple but powerful self-correction method for the Min-Sum decoding of LPDC codes. Unlike other correction methods known in the literature, our method does not try to correct the check node processing approximation, but it modifies the variable node processing by erasing unreliable messages. However, this positively affects check node messages, which become symmetric Gaussian distributed, and we show that this is sufficient to ensure a quasi-optimal decoding performance. Monte-Carlo simulations show that the proposed Self-Corrected Min-Sum decoding performs very close to the Sum-Product decoding, while preserving the main features of the Min-Sum decoding, that is low complexity and independence with respect to noise variance estimation errors.