Dynamic Spectrum Management Project



Channel Identification Methods

One major impairment of current DSL systems is severe crosstalk between telephone lines that occupy the same cable or conduit. Such crosstalk limits the reliability and/or maximum data rates of current DSL services. Crosstalk also degrades the existing services when a new service is added to the current bundle of lines within a cable. To avoid DSL service degradation, current systems use a 'worst-case crosstalk scenario'. In order to enable dynamic management of resources, it is essential to develop methods to measure the current crosstalk power spectra and signals with reasonable accuracy. Important channel parameters to identify can include line transfer functions, non-crosstalk noise power levels, transmit power levels and DSL transmitters' information. Better knowledge of crosstalk, along with these channel parameters, can allow service providers to:

  1. Predict data rates and ranges in DSL systems with better accuracy
  2. Predict and isolate service failure to degrading channel conditions (including assistance in pair selection)
  3. Create different user service tiers by varying user rates and symmetries
  4. Improve performance greatly by using the channel data for dynamic spectrum management

For more on this please refer to our T1E1.4 contribution, "Channel Identification Proposal (147)" (ms word format). Also refer to our other T1E1.4 contributions.

Our journal and conference publications on Channel Identification Methods give further details on methods that we are working on in this field.
 


http://cafe.stanford.edu/~cioffi/dsm/