Reduced Complexity Neural Network Equalizers for Two-Dimensional Magnetic Recording

This article investigates the reduced complexity neural network (NN)-based architectures for equalization over two-dimensional magnetic recording (TDMR) digital communication channel data storage. We use realistic waveforms measured from a hard disk drive (HDD) with TDMR technology. show that multilayer perceptron (MLP) nonlinear equalizer achieves 10.91% reduction in bit error rate (BER) linear cross-entropy (CE)-based optimization. However, MLP equalizer’s is $6.6\times $ complexity. Thus, we propose (RC-MLP) equalizers. Each RC-MLP variant consists of finite-impulse response (FIR) filters, activation, and hidden delay line. A proposed entails only notation="LaTeX">$1.59\times while achieving 8.23% BER equalizer.