ADAPTIVE IDENTIFICATION AND EQUALIZATION OF MAGNETIC RECORDING CHANNELS

A new RLS adaptive Volterra filter is presented. The nonlinear filteri ng problem is transformed into an equivalent multichannel, but linear, filtering problem. The multichannel input signal is completely orthog onalized using sequential processing multichannel lattice stages. Thus , a fast convergent, highly modular and, simple filter with good numer ical properties is designed. In the identification of magnetic recordi ng channels, the filter identifies the channels directly and parameter s for the channel nonlinearity are quantified simultaneously. In the e qualization of magnetic channels, the most effective equalizer length can be assigned dynamically.