Nonlinear equalization for reduction of nonlinear distortion in high-density recording channels

Abstract

Three different types of structures for nonlinear equalizers are discussed to reduce nonlinear distortion in digital high-density recording channels. The first two equalizers, a nonlinear DFE and a modified version of partial response maximum likelihood (PRML) with a nonlinear equalizer, which are based on the pth-order inverse theorem of Volterra series, can reduce the nonlinear ISI existing in both precursor and postcursor parts. The last one, a nonlinear canceller with an appropriate tentative decision device, can further cancel the cross products between the previous and future symbols without producing higher-order terms (/spl ges/p+1). By comparing their performances via simulation with the existing LE, linear DFE, RAM-DFE, and PRML for digital recording channels with three different densities, the authors confirm that the nonlinear ISI can be reduced effectively by the approaches discussed in the paper especially at higher densities.