Proactive Fiber Damage Detection in Real-time Coherent Receiver

Abstract

We develop an algorithm extension for a coherent receiver, coupled with machine learning to monitor mechanical stress to an optical fiber, for recognizing fiber breaks before they occur. We demonstrate event classification with 95% accuracy over a real-time PDM-QPSK testbed.