Improving DML-Based OFDM Transmission by a Filter-Aided Neural Network Equalizer

Abstract

We propose a filter-aided NN to compensate 10G-DML-based OFDM transmission, avoiding low-frequency SNR degradation. After 100- and 150-km SMF, the data rates are increased by 95 and 130% to achieve ∼40 and 32 Gbps, respectively.