A Detection Scheme With TMR Estimation Based on Multi-Layer Perceptrons for Bit Patterned Media Recording

Abstract

Increasing areal density in bit patterned media recording systems increases the 2-D interference from the adjacent islands in the down-track and cross-track directions. Also, nonlinear effects due to the track mis-registration (TMR) and media noise become severe as the areal density increases and degrade the bit error rate (BER) performance. In this paper, we proposed a new detection scheme based on multi-layer perceptrons (MLPs) to alleviate the nonlinear effects. Two MLPs are used in the proposed detection scheme and learn the channel with the nonlinear effects in the training procedure. One MLP is used for TMR estimation and the other MLP is used for data detection. The MLP-based TMR estimator estimates the TMR from the readback signals. Then, the MLP-based data detector equalizes the channel reflecting the estimated TMR and media noise and detects the data bit. Therefore, the detection scheme is robust to the effects of the TMR and media noise. The proposed detection scheme provides the improved BER performances compared to the conventional detection scheme using the 2-D equalizer filter, 1-D generalized partial response target, and Viterbi detector in the channel with the TMR and media noise.