Track Misregistration Estimation Technique Based on Hybrid <i>K</i> -Means and EM Algorithm in Bit-Patterned Media Recording Systems

Abstract

Track misregistration (TMR) in ultrahigh density bit-patterned media recording (BPMR) is a significant issue, severely degrading system performance. Although TMR can be managed by a servo control loop, this article proposes a hybrid TMR mitigation method based on K-means and expectation-maximization (EM) algorithms to enhance TMR prediction accuracy and improve bit-error-rate (BER) in multihead/multitrack BPMR systems. This method utilizes 2-D equalizer and 1-D generalized partial response (GPR) target coefficients for the soft-output Viterbi algorithm (SOVA) detector according to the estimated TMR level to mitigate this effect. Simulation results demonstrate that the proposed system significantly outperforms conventional systems, especially under high TMR conditions. The hybrid approach achieves high TMR estimation accuracy and delivers BER performance close to an ideal system with perfect TMR estimation, showing up to a 1.25 dB improvement in BER over systems without TMR mitigation. These findings underscore the effectiveness of the hybrid K-means-EM-based TMR estimator in enhancing system performance under various conditions.