Multilayer Perceptron-Based Soft-Information Modification Technique for Double-Layer Bit-Patterned Magnetic Recording Systems

Abstract

The growing demand for data storage necessitates continuous improvements in hard disk drive (HDD) technologies. Double-layer magnetic recording (DLMR) and bit-patterned magnetic recording (BPMR) technologies are pivotal in enhancing areal density (AD), but they also introduce challenges such as inter-layer interference (ILl) and two-dimensional (2D) inter-ference. This study addresses these challenges by integrating a multilayer perceptron (MLP)-based soft-information modifier into the double-layer BPMR system. The proposed method optimizes the log-likelihood ratio sequence for the recorded tracks on the second recording layer, which are more prone to interference compared with the first recording layer. Our simulations demonstrate significant improvements in bit-error-rate performance across varying levels of media noise and track misregistration. The results also indicate that the MLP-based approach effectively mitigates the adverse effects of interference, thereby enhancing the reliability of data retrieval in high-AD HDD systems.