Developing A Fault Diagnosis Approach Based On Artificial Neural Network And Self Organization Map For Occurred ADSL Faults

Telecommunication companies have received a great deal of research attention, which have many advantages such as low cost, higher qualification, simple installation and maintenance, and high reliability. However, the using of technical maintenance approaches in Telecommunication companies could improve system reliability and users' satisfaction from Asymmetric digital subscriber line (ADSL) services. In ADSL systems, there are many variables giving some noise for classification and there are many fault patterns with overlapping data. Therefore, this paper proposes a multilayer perceptron (MLP) classifier integrated with Self Organization Map (SOM) models for fault detection and diagnosis (FDD) of occurred ADSL systems. The interest of this paper is to improve the performance of single MLP by dividing the fault pattern space into a few smaller sub-spaces using SOM clustering technique and triggering the right local classifier by designing a supervisor agent. The performances of this method are evaluated on the fault data of Iranian Telecommunication Company which develop ADSL services and then the proposed algorithm is also compared against single MLP. Finally, the results obtained by this algorithm are analyzed to increase user's satisfaction with reducing occurred faults for them with predicting before they face it.