DETECTION FOR CHANNELS WITH TRANSITION NOISE

Transition noise is known to be a major cause of errors for high densi ty magnetic recordings This noise is signal dependent and can be model ed as multiplicative noise in a linear channel model. Maximum-likeliho od method was not considered for detection of signals in such noise in the past. In this study, a detector model for an asymptotic maximum-l ikelihood (AML) detection is developed for systems with such noise. Ba sed on a linear partial response channel model, a recursive procedure is obtained as a tree search algorithm, leading to the maximum likelih ood detection asymptotically, as the tree-search depth is increased. P erformance estimation will be discussed in a separate paper.