Non-linearities mitigation with fuzzy neural networks using a machine learning algorithm in a CO-OFDM system

摘要

This study presents a fuzzy neural network based non-linear equaliser to diminish the non-linearities in a coherent optical orthogonal frequency division multiplexing (CO-OFDM) system. The numerical results show that the proposed technique based CO-OFDM system outperforms the CO-OFDM system without non-linear equaliser by 5 and 7.5% EVM performance, and 2.36 and 4.87 dB Q-factor performance after 1000 km transmission and -3 dBm input launch power at a bit rate of 40 and 80 Gbps, respectively. Moreover, it has been generally accepted in statistics that the rank-based Wilcoxon methodology provide more robust results in a contradiction of outliers. Therefore, the aim of this study is to analyse fuzzy neural network based non-linear equaliser and compare the results with that of Wilcoxon approach fuzzy neural network based non-linear equaliser.