Post Equalization Scheme Based on Deep Neural Network for a Probabilistic Shaping 128 QAM DFT-S OFDM Signal in Underwater Visible Light Communication System

Abstract

We have presented a post equalization scheme based on Deep Neural Network (DNN) for DFT-S OFDM modulation using the probabilistic shaping (PS) technique in underwater visible light communication (VLC) system. By this method, we successfully demonstrated a data rate of 1.74Gbit/s PS128QAM DFT-S OFDM modulation over 1.2meter underwater optical transmission with bit error rate (BER) below 7% FEC threshold of 3.8×10-3. Compared to the typical PS128QAM DFT-S OFDM modulation without DNN, the proposed method would lead to an improvement of system capacity of 5.4% by increasing the data rate by 90 Mbps. The experimental results validate that the proposed DNN-based post equalization scheme for odd order QAM PS technique can be a promising solution for future high speed underwater VLC system.