Track Mis-Registration Estimator Based on K-Means Algorithm for Bit-Patterned Media Recording

Bit-patterned media recording (BPMR) has been considered among the key technologies to extend recording densities to 1 Tb/in2 and beyond. However, the BPMR system encounters the challenges of two-dimensional intersymbol interference and track mis-registration (TMR). In this article, we propose a TMR estimator based on the $K$ -means algorithm for estimating the TMR levels and preventing the performance degradation of the BPMR system. The TMR estimator calculates the distances between the centroid of the readback signal and predetermined centroids according to the TMR level. The equalizer coefficients and partial response target are adjusted to the corresponding TMR level. At a bit error rate of 10−3 and a TMR of 10%, the proposed detection scheme performs better by approximately 1 dB in comparison with the conventional detection.

I. Introduction

As the volume of digital data created by people and devices continues to grow, the demand for higher recording density in magnetic recording systems has increased [1]. However, conventional magnetic recording systems face a recording density limit of 1 Tb/in² due to the superparamagnetic effect, which leads to data loss. To obtain an increased areal density (AD) of 1 Tb/in² and greater, bit-patterned media recording (BPMR) has been considered among the promising future magnetic storage systems [1]. Because a single magnetic island can store 1 bit and these islands are surrounded by nonmagnetic regions in the BPMR system, BPMR presents advantages, such as reduced nonlinear transition shift and transition noise, as well as simplified tracking [2]. In traditional magnetic recording systems where data are organized along single track, the interaction between neighboring tracks during writing and reading of data can be neglected and data storage channel has been considered as one-dimensional (1-D) interference. However, the conventional storage systems face two-dimensional (2-D) interference because the intertrack distance should be decreased for higher AD.