Crosstalk prediction on wires bundles by Kriging approach
by
Nadjet Dahim
CNRS-SUPELEC
Coauthors: Walid Tabbara (Laboratoire CNRS-Supelec)

It deals with the crosstalk prediction of complex bundles of wires. Kriging approach is introduced as a very efficient mutifactor prediction method that allows to avoid the high computational cost of transmission line simulations and that provides designer with real time answers in order to find out the best configuration that minimizes the crosstalk. Kriging predictor is suitable for this kind of problems because it provides good accuracy even from a little amount of information. This approach combines a regression model with a stochastic process (a Gaussian type had been chosen). It provides the Best Linear Unbiased Predictor of an observable, and the uncertainty of the prediction in term of the mean squared error.

Date received: April 14, 2000

Copyright © 2000 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cadt-03.