Serial Detection with Neural Network-Based Noise Prediction for Bit-Patterned Media Recording Systems

Ultra-high density data storage has gained high significance given the increasing amounts of data; many technologies have been proposed to achieve a density. Among them, bit-pattern media recording (BPMR) is promising technology. In BPMR systems, are stored on magnetic islands. Therefore, densities can be achieved by reducing distance between Because closeness islands, readback signal distorted two-dimensional (2D) interference, which includes intersymbol interference according down-track direction and intertrack cross-track direction. A simple effective serial detection algorithm was recently mitigate 2D interference. However, utilizes hard output in inner detection, this degrades performance. To resolve problem, subsequent study used feedback estimate noise create soft for detection. Following up, paper we propose model that neural network prediction. The network-based with line were compared terms bit error rate (BER). results show achieves gain approximately 1 dB at BER 10−6.