Comparison Between Viterbi Detectors for Magnetic Recording Channels Based on Regressive and Autoregressive Noise Models

Abstract

A recent work presents a regressive noise model for the data-dependent correlated noise, at the output of a magnetic recording channel detector. We have generalized this channel model, considering digital equalization and a more efficient correlation matrix, in order to make a comparison with the usual detector in a more realistic environment. Simulation results show that the regressive detector performs better when the number of trellis states is lower than needed, while both approaches are comparable when the number of states matches the channel memory.