The Digital Subscriber Line (DSL) systems performance tightly depends on noise interference. The users (lines) in the binder create mutual interference (crosstalk), therefore decreasing the rates of all users. Crosstalk noise is more predominant in DSL and its major performance bottleneck is to limit high-speed data rate. This study proposes a new real-time monitoring methodology for noise estimation based on Management Information Base (MIB) metrics. Linear regression is used for fitting, in which input parameters are MIBs and the output is the estimated noise power. The results confirm the possibility of estimating noise with a general equation. Moreover, it is showing the result improvement if a Loop Topology Identification tool is used as prior knowledge.
Resumo-Este artigo propõe uma nova metodologia para identificação de ruídos em redes Very-high-bit-rate Digital Subscriber Line (VDSL). Uma Rede Neural Artificial (RNA) baseada no algoritmo backpropagation é usada para classificar, em tempo real, a ausência de ruído ou a predominância de ruído impulsivo, ou crosstalk. Os resultados demonstraram o potencial da utilização de algoritmos de inteligência computacional para a identificação de ruídos em redes VDSL. Dentre os testes, o melhor resultado alcançado foi de 99.3% de acerto.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.