Branch Identification in Passive Optical Networks using Machine Learning
arXiv:2304.00285
Abstract
A machine learning approach for improving monitoring in passive optical networks with almost equidistant branches is proposed and experimentally validated. It achieves a high diagnostic accuracy of 98.7% and an event localization error of 0.5m
Optical Fiber Communication Conference (OFC) 2023