Simplified Variable-Scaled Min Sum LDPC decoder for irregular LDPC Codes

TL;DR

This paper introduces a simplified, variable-scaled Min-Sum LDPC decoder that adapts the scaling factor during iterations, significantly improving decoding performance with minimal added complexity.

Contribution

It proposes a novel staircase-based variable scaling factor for Min-Sum decoding, enhancing performance while maintaining implementation simplicity.

Findings

Performance improved by variable scaling factor

Staircase design optimizes storage and performance

Simulation confirms non-trivial decoding gains

Abstract

Min-Sum decoding is widely used for decoding LDPC codes in many modern digital video broadcasting decoding due to its relative low complexity and robustness against quantization error. However, the suboptimal performance of the Min-Sum affects the integrated performance of wireless receivers. In this paper, we present the idea of adapting the scaling factor of the Min-Sum decoder with iterations through a simple approximation. For the ease of implementation the scaling factor can be changed in a staircase fashion. The stair step is designed to optimize the decoder performance and the required storage for its different values. The variable scaling factor proposed algorithm produces a non-trivial improvement of the performance of the Min-Sum decoding as verified by simulation results.