On Implementation of Min-Sum Algorithm for Decoding Low-Density Parity-Check (LDPC) Codes

摘要

This paper is concerned with the implementation issues of the so-called rain-sum algorithm (also referred to as max-sum or max-product) for the decoding of low-density parity-check (LDPC) codes. The effects of clipping threshold and the number of quantization bits on the performance of min-sum algorithm at short block lengths are studied. It is shown that min-sum is robust against quantization effects, and in many cases, only four quantization bits suffices to obtain close to ideal performance. We also propose modifications to min-sum algorithm that improve the performance by a few tenths of a dB with just a small increase in decoding complexity.