Model-Based Detection of Anomalous Events in Submarine Cables Using Distributed Deformation Sensing and Kalman Filtering
For operators of submarine cable infrastructure, this provides an interpretable anomaly detection method with controlled false-alarm rates, though it is an incremental application of existing techniques to a specific domain.
This paper proposes a model-based framework for real-time anomaly detection in submarine cables using distributed deformation sensing and Kalman filtering, demonstrating reliable identification of localized disturbances while remaining robust to ambient excitation in numerical simulations.
Submarine power and telecommunication cables constitute critical global infrastructure, yet they remain vulnerable to mechanical damage caused by maritime activities and intentional tampering. Continuous monitoring of these assets is therefore essential for early detection of anomalous events. This paper proposes a model-based framework for real-time anomaly detection in submarine cables using spatially distributed deformation measurements along the cable. The cable is modeled as a tensioned structure governed by a damped wave equation with fixed boundary conditions. A finite-dimensional state-space representation is obtained through spatial discretization, enabling the use of a Kalman filter to estimate the cable's dynamic state under stochastic environmental disturbances. Anomaly detection is then formulated as a statistical hypothesis test applied to the innovation sequence of the filter. Compared with purely data-driven alarms, the proposed framework provides an interpretable residual signal whose threshold can be related to a prescribed false-alarm probability. Numerical simulations demonstrate that the proposed framework can reliably identify localized disturbances while remaining robust to ambient environmental excitation.