Finite alphabet iterative decoders robust to faulty hardware: Analysis and selection

In this paper, we analyze Finite Alphabet Iterative Decoders (FAIDs) running on faulty hardware. Under symmetric error models at the message level, we derive the noisy Density Evolution equations, and introduce a new noisy threshold phenomenon (called functional threshold), which accurately characte...

Full description

Saved in:
Bibliographic Details
Main Authors: Dupraz, Elsa, Declercq, David, Vasic, Bane, Savin, Valentin
Format: Conference Proceeding
Language:English
Subjects:
Decoding
Error probability
Hardware
Iterative decoding
Noise
Noise measurement
Robustness
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we analyze Finite Alphabet Iterative Decoders (FAIDs) running on faulty hardware. Under symmetric error models at the message level, we derive the noisy Density Evolution equations, and introduce a new noisy threshold phenomenon (called functional threshold), which accurately characterizes the convergence behavior of LDPC code ensembles under noisy-FAID decoding. The proposed functional threshold is then used to identify FAIDs which are particularly robust to the transient hardware noise. Finite-length simulations are drawn to verify the validity of the asymptotical study.
ISSN:2165-4700
DOI:10.1109/ISTC.2014.6955095