Use of the neural networks in the qualification of the XDSL loops Internet services in Etecsa

Abstract

The last years have been characterized by enormous efforts in the development of high speed transmission technologies through local telephone network copper pairs, known as xDSL (HDSL, ADSL, VDSL, etc.). This article deals with the problem related to the physical parameter influence of copper pairs measured by the Telephone exchange for ADSL —Asymmetric Digital Subscriber Line— services and the use of Artificial Neural networks for the estimation prediction of the maximum Theoretical speed to be achieved for each subscriber, in particular taking into account the maximum number of subscribers that can simultaneously coexist on the same cable and the measurement of the telephone model parameters, with a successful result implemented with the BinteK software.