Recovery of ISI channels using multiresolution wavelet equalization

Equalization of ISI is usually performed in the time domain. In fact it is difficult to operate in the frequency domain, since with the usual spectral techniques based on the Fourier transform, correspondence between temporal and frequency components is lost in the transformation. In this paper we present an application of the wavelet transform that exploits the joint time-frequency representation of this transform to equalize channels affected by ISI. It is shown that using the description of wavelets as filters, and exploiting the multispectral decomposition of the received signal, ISI can be dramatically reduced.