Improved Multitrack Detection With Hybrid 2-D Equalizer and Modified Viterbi Detector

Abstract

Increasing track density in magnetic recording systems increases inter-track interference (ITI). One way to handle the increased ITI is to apply 2-D equalization to the signals from multiple tracks. Usually the equalizer coefficients and the corresponding partial response (PR) targets are fixed after training using a pseudo-random bit sequence. We denote such an equalizer as a fixed equalizer. In this paper, we propose and investigate a variable 2-D equalizer for bit patterned media recording (BPMR). In the proposed approach, the equalizer coefficients and corresponding PR targets vary with ITI patterns. These variable equalizer coefficients are determined and stored by training with sequences representing different ITI patterns. In the proposed approach for BPMR, during data detection, these variable equalizer coefficients and corresponding PR targets are embedded into the parallel branches of a modified Viterbi (MV) detector in response to estimated ITI patterns. We term this new approach as V2DEMV [i.e., variable 2-D equalizer with 2-D PR target followed by MV detector] algorithm. Furthermore, we propose to simplify the implementation of the V2DEMV algorithm by identifying the survivor branch among the branches corresponding to the dominant erroneous ITI patterns. Then this survivor branch and other branches corresponding to remaining possible ITI patterns are processed by an F2DEMV (i.e., fixed 2-D equalizer with 2-D PR target followed by MV detector) algorithm to estimate the bits. We term this reduced-complexity approach as an H2DEMV (i.e., hybrid 2-D equalizer with 2-D PR target followed by MV detector) algorithm. Our simulations indicate that for both two readers detecting two tracks and three readers detecting three tracks, at target bit error rate of 10 <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-3</sup> , the proposed H2DEMV algorithm can provide noticeable signal-to-noise ratio (SNR) gains compared to F2DE1DV algorithm (i.e., fixed 2-D equalizer with 1-D PR target followed by 1-D Viterbi detector) and F2DEMV algorithm.