Near-end crosstalk mitigation using wavelets

A new method to mitigate near-end crosstalk (NEXT) in the wavelet transform domain is proposed. The method utilizes the fact that the received signal has greater energy in the lower end of the transmission bandwidth in comparison to the NEXT noise and therefore is more regular, or smooth, than the NEXT noise. The method entails a new technique of estimating the NEXT noise from the received signal so that appropriate denoising methods can be applied to reduce the NEXT noise. Experiment results have shown that the method is quite effective in reducing NEXT noise especially when the signal to crosstalk noise ratio is low (SCNR). Furthermore, the method has a low computational complexity which makes it fast, efficient, and well suited for high datarate applications.