Impulse noise identification for the HFC upstream channel

Abstract

An algorithm is described that identifies the shape of the impulse within the impulse noise component of the HFC upstream channel. The cross-correlation function (CCF) is used to align samples of the impulse waveforms in time. The singular value decomposition (SVD) finds the dominant basis waveforms that represent the collection of samples. We find that two frequency-modulated basis waveforms sufficiently span our set of 75 measured impulses. The joint statistics of the basis coefficients are given.