Identificação e estimação de ruído em redes DSL: uma abordagem baseada em inteligência computacional

Abstract

This paper proposes the use of computational intelligence techniques aiming to identify and estimate the noise power in Digital Subscriber Line (DSL) networks on real time. A methodology based on Knowledge Discovery in Databases (KDD) for detect and estimate noise in real time, was used. KDD is applied to select, pre-process and transform data before data mining step. For noise identification the traditional backpropagation algorithm based on Artificial Neural Networks (ANN) is applied aiming to identify the predominant noise during the collection of information from the user's modem and the DSL Access Multiplexer (DSLAM). While the algorithm for noise estimation, linear regression and a hybrid algorithm consisting of Fuzzy with linear regression are applied to estimate the noise power in Watts. Results show that the use of computational intelligence algorithms such as RNA are promising for noise identification in DSL networks, and algorithms such as linear regression and fuzzy with linear regression (FRL) are promising for noise estimation in DSL networks.