SELT measurements have proven to be useful for acquiring information on the make-up of a single DSL loop. Loops that run through the same binder may see common features in their SELT data, separate from the distinct features that indicate the end of the line. For the moment, this information is not used in the topology extraction of DSL loops. At the binder level we may expect a higher correlation in the features of lines pertaining to the same binder. The aim is to model this joint reaction and correlation, which would be useful to extract more information beyond the per-line state-of-the art.
Moreover, binder features are typically detected after converting the frequency domain SELT-PMD to the time domain [Ref-Diamond]. Working in the time domain has several disadvantages, such as high signal dispersion, low measurement resolution and conversion losses due to the use of suboptimal band-pass filters. It will be explored how these challenges can be overcome by applying additional processing. One possibility is to add additional processing in the frequency domain, or to shift parts of the processing from the time domain to the frequency domain. Ultimately, the signal processing used to extract the topology features (per-line and/or per-binder) can be carried out entirely in the frequency domain, which may require the development of new frequency domain models.
Effective start/end date1/02/0831/01/10

    Research areas

  • Automatic Measurement Systems, Nonlinear Modelling, Medical Physics, Fibre Optic, Parameter Estimation, Microwaves, Instrumentation, Underwater Acoustics, System Identification, Telecommunications, Electrical Measurements, Nonlinear Measurements, Electromagnetism

    Flemish discipline codes

  • Electrical and electronic engineering
  • Mathematical sciences
  • (Bio)medical engineering

ID: 3260388