Noise Estimation Proposal for Real Time DSL Systems using Linear Regression and Fuzzy Systems

This paper presents two methods for determining the noise power (mainly caused by crosstalk) in Digital Subscriber Line (DSL) networks. A fuzzy system approach is compared to a linear regression approach. Both are applied to a real world DSL network where a variable noise power is injected. Knowledge Discovery in Database (KDD) is used to organize the data selection, while linear regression and Fuzzy algorithms are compared through data mining process. Results show that the use of techniques such as Fuzzy and linear regression are an effective solution for the real-time estimation of crosstalk in DSL systems. The proposed scheme can be adapted to other types of noise, thus extending its application to DSL systems.