DMCA
Some High-Rate Near Capacity Codecs for the Gaussian Channel (1996)
| Venue: | 34th Allerton Conference on Communications, Control and Computing |
| Citations: | 23 - 2 self |
BibTeX
@INPROCEEDINGS{Cheng96somehigh-rate,
author = {Jung-fu Cheng and Robert J. Mceliece},
title = {Some High-Rate Near Capacity Codecs for the Gaussian Channel},
booktitle = {34th Allerton Conference on Communications, Control and Computing},
year = {1996}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We present a construction of high-rate linear error-correcting codes based on random bipartite graphs. In contrast to other researchers, our approach uses these graphs to produce generator matrices. A distributive parallel decoding algorithm based on the principle of belief propagation is presented. These new codes are then shown to be linear-time encodable and decodable. Experimental results further show the performances of these new systems to be approach the channel capacities of the Gaussian channels with binary or multilevel signal constellations. 1 Introduction The introduction of multi-stage iterative decoding with exchange of soft information [3] has recently attracted substantial research interests and reperesents a genuine breakthrough in coding theory. This methodology has been applied to parallel concatenated convolutional (turbo) codes [5, 11, 12, 17, 22], parallel concatenated block codes [17], serial concatenated convolutional codes [2], and generalized concatenated con...
Keyphrases
gaussian channel high-rate near capacity codecs genuine breakthrough soft information new system block code new code high-rate linear error-correcting code substantial research interest belief propagation distributive parallel random bipartite graph convolutional code experimental result channel capacity generator matrix multilevel signal constellation
