Presentation Information

[WF1-4]ML-Aided Proactive In-line EDFAs’ Gain Degradation Detection and Localization in Optical Networks

○Hongcheng Wu1, Qi Hu1, Zhuojun Cai1, Gai Zhou2, Kangping Zhong3, Faisal Nadeem Khan1 (1.Tsinghua University, 2.Guangdong University of Technology, 3.The Hong Kong Polytechnic University)

Keywords:

Artificial intelligence and machine learning for optical network design,control,and management

We propose a machine learning-based framework inside standard coherent receivers for proactive detection and localization of in-line EDFAs’ gain degradation. 84.2% failures can be detected in advance and 93.1% failures localization accuracy is experimentally demonstrated.