Joint zero-forcing and matched filter processing in an adaptive equalizer using the linearly-constrained least-mean square (LC-LMS) algorithm

Digital receivers often contain a cascade of two filters in the base-band signal-conditioning path. The first is a square-root raised-cosine Nyquist matched filter designed to maximize SNR and suppress out-of-band interference, while the other is a decision-directed (T/2)-spaced equalizer which removes the spectral distortion introduced by the channel. This paper presents an equalizer modified to permit a single filter to converge simultaneously to a (T/2)-spaced zero forcing equalizer and a matched filter.