Fiber Optics for AI-based Cable Threat Detection in Urban and Submarine Environments

Protecting critical power infrastructure requires reliable detection of activities that threaten cables in urban and submarine environments. Distributed acoustic sensing (DAS) enables continuous monitoring by capturing vibrations along the cable route. In this work, we analyze signal characteristics for activities that can lead to cable damage, including excavator digging, jackhammering, auger drilling, and diver interference. Based on this analysis, we present a detection framework that combines deep learning and signal processing to enhance adaptability and robustness.