Mitigating Track Misregistration Using the DBSCAN Algorithm for Single-Reader/Two-Track Reading in Shingled Magnetic Recording Systems

Abstract

Shingled magnetic recording (SMR) systems can effectively reduce track width using the shingled writing technique, resulting in significantly higher areal density than conventional magnetic recording. However, track misregistration (TMR) still frequently occurs, leading to read errors and reduced signal processing performance. To address this, we propose a method for estimating TMR levels for single-reader/two-track reading in the SMR system using a density-based spatial clustering of applications with noise (DBSCAN) algorithm. Furthermore, we also present a mitigation method for TMR effects using the DBSCAN algorithm. In the TMR-level estimation process, the equalized signal from the first equalizer is fed into the first DBSCAN algorithm. The estimated TMR level is then used to select the appropriate equalizer to equalize the readback signal. Finally, the DBSCAN-based detector is used to detect the equalized signal. Simulation results at an areal density of 2 Tb/in2 demonstrate that our proposed method can accurately predict TMR levels, effectively mitigate TMR effects, and improve overall recording performance in terms of bit-error rate.