deposit_your_work

Blind Crosstalk Cancellation for DMT Systems

Files in this item

Files Size Format View
Ahm2002Sep5BlindCross.PDF 198.0Kb application/pdf Thumbnail

Show full item record

Item Metadata

Title: Blind Crosstalk Cancellation for DMT Systems
Author: Ahmed, Nadeem; Warke, Nirmal; Baraniuk, Richard G.
Type: Conference Paper
Keywords: crosstalk; NEXT; mitigation; cancellation
Citation: N. Ahmed, N. Warke and R. G. Baraniuk,"Blind Crosstalk Cancellation for DMT Systems," in IEEE ETTC,
Abstract: As the deployment of broadband communication systems such as DSL continues to grow, system performance in terms of capacity and error rates is severely limited by crosstalk interference. In order to continue deploying high speed DSL service, it becomes necessary to mitigate this crosstalk interference. All current crosstalk mitigation techniques require precise knowledge of the crosstalk coupling functions between the twisted pair wires carrying DSL service within a cable binder. In practice, this is near impossible to obtain, without coordination between the different DSL services. In this paper, we present a crosstalk cancellation technique that is blind to the coupling functions between the wire pairs. Our technique uses the statistical properties of the received crosstalk signal (that can be easily estimated), rather than the exact coupling functions themselves, making it very attractive for practical implementation. We show with the help of simulations for a realistic ADSL system with HDSL and T1 crosstalk interference using actual measured crosstalk coupling functions that the proposed blind crosstalk cancellation technique can achieve significant gains in terms of rate and reach improvement for the ADSL system. We also show that the performance of the proposed blind crosstalk cancellation technique is robust to a jitter in the crosstalk symbol timing estimate that is required to construct the cancellation signal.
Date Published: 2002-09-01

This item appears in the following Collection(s)

  • ECE Publications [1045 items]
    Publications by Rice University Electrical and Computer Engineering faculty and graduate students
  • DSP Publications [508 items]
    Publications by Rice Faculty and graduate students in digital signal processing.