CODED 64-CAP ADSL IN AN IMPULSE-NOISE ENVIRONMENT - MODELING OF IMPULSE NOISE AND FIRST SIMULATION RESULTS

Citation
W. Henkel et al., CODED 64-CAP ADSL IN AN IMPULSE-NOISE ENVIRONMENT - MODELING OF IMPULSE NOISE AND FIRST SIMULATION RESULTS, IEEE journal on selected areas in communications, 13(9), 1995, pp. 1611-1621
Citations number
25
Categorie Soggetti
Telecommunications,"Engineering, Eletrical & Electronic
ISSN journal
07338716
Volume
13
Issue
9
Year of publication
1995
Pages
1611 - 1621
Database
ISI
SICI code
0733-8716(1995)13:9<1611:C6AIAI>2.0.ZU;2-3
Abstract
This paper presents the performance of various coding schemes for the asymmetrical digital subscriber line (ADSL) in an impulse-noise enviro nment, Impulse noise is considered to be one of the most damaging impa irments in the ADSL, in which compressed video signals are delivered t o residential customers. The impulse noise used in this study was meas ured and collected in German telephone networks. Based on this measure ment and the corresponding statistical modeling, a simulation model fo r impulse noise is proposed and its properties are outlined. The codin g schemes considered here utilize burst-error correcting Reed-Solomon codes and/or random error correcting trellis codes as well as symbol i nterleaving between the two codes. It has been found through computer simulations that a proper concatenation of the two codes could increas e the immunity against impulse noise compared to an uncoded scheme. Sp ecifically, a concatenated code, using a 2-dimensional 8-state trellis code and a 4-error-correcting Reed-Solomon code with an interleaving depth of 18 symbols, was able to eliminate all the errors caused by th e impulse noise used in the study. It has also been found that the tre llis codes are not very effective against impulse noise, unless they a re used in conjunction with Reed-Solomon codes and a proper symbol int erleaving. Performance results of other coding configurations using Re ed-Solomon codes with different error-correcting capabilities are also presented. In addition, we also show the performance results when sim ple array codes are used instead of the Reed-Solomon codes.