A model based approach to fault detection for the reverse path of cable television networks

摘要

We present a model based method for reliably detecting faults in the reverse path of cable amplifier networks. This method has the advantage over traditional fixed-bound fault detection techniques in that it is able to accurately detect changes in signal behaviour while tracking signal changes due to environmental effects. The resulting method provides an increase in the fault detection sensitivity while simultaneously providing a decrease in the false alarm rate. We have implemented a general approach based an using a modeling engine to capture the reverse pilot signal behaviour of cable television amplifiers. Two modeling specific engines were developed for this purpose. The first one is based on the use of feedforward neural networks; the second one is based on the use of statistical analysis techniques. The resulting fault detection system, when employing either modeling engine, was able to provide good temporal localization of the onset of fault conditions along with a clear indication of the presence of the fault through its duration.