Detection and Classification of Eavesdropping and Mechanical Vibrations in Fiber Optical Networks by Analyzing Polarization Signatures Over a Noisy Environment

We propose a machine-learning-based method to detect and classify eavesdropping and mechanical vibrations in an optical network based on state of polarization variations. Tests in two realworld installations with links of different lengths demonstrate an accuracy of 86.5% in 7 distinct normal and malicious scenarios.