Title :
On LDPC codes over symmetric channels
Author :
Saeedi, Hamid ; Pishro-Nik, Hossein
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Massachusetts, Amherst, MA, USA
Abstract :
In the past decade, there has been tremendous amount of research on the analysis and design of the Low-Density Parity-Check (LDPC) codes with belief propagation decoding over different types of Binary-Input Output-Symmetric Memoryless (BIOSM) channels. However, with the exception of the Binary Erasure Channel (BEC), analytical results on LDPC codes over such channels are limited and most results are based on numerical methods and optimization. In particular, systematic design of provably capacity achieving sequences of LDPC code ensembles over general class of BIOSM channels has remained a fundamental open problem. Such sequences have been designed only for the binary erasure channel (BEC). In this paper, we first prove some novel analytical properties of the LDPC code ensembles over BIOSM channels. In particular, we prove a result that generalizes the previously known stability condition over symmetric channels. This suggests that a modified version of the flatness condition, a property which has been shown to be critical for capacity achieving sequences over the BEC, can be used to devise capacity achieving sequences for general class of BIOSM channels. Based on this assumption, we propose a method which could result in the systematic design of such sequences over BIOSM channels. Numerical evidence is promising and provides a consistent convergence behavior to capacity over the considered BIOSM channels as the average check node degree increases.
Keywords :
channel coding; decoding; numerical analysis; optimisation; parity check codes; LDPC codes; average check node degree; belief propagation decoding; binary erasure channel; binary-input output-symmetric memoryless channel; channel coding; low-density parity-check code; numerical method; optimization; symmetric channel; Additive white noise; Belief propagation; Convergence of numerical methods; Decoding; H infinity control; Optimization methods; Parity check codes; Performance analysis; Stability analysis; Upper bound;
Conference_Titel :
Networking and Information Theory, 2009. ITW 2009. IEEE Information Theory Workshop on
Conference_Location :
Volos
Print_ISBN :
978-1-4244-4535-6
Electronic_ISBN :
978-1-4244-4536-3
DOI :
10.1109/ITWNIT.2009.5158547