Kalman filter detects damage in submarine cables with 99% accuracy

Submarine cables—backbone of global internet and power grids—are vulnerable to ship anchors, fishing trawlers, and deliberate sabotage. Current monitoring often relies on data-driven methods that lack transparency. A team led by Gabriele Oliva at Università Campus Bio-Medico di Roma (with colleagues from University of Rome Tor Vergata) published a paper accepted for the MED2026 conference that introduces a model-based framework to catch anomalies in real time.

The approach models the cable as a damped wave equation with fixed ends, discretized into a finite state-space. A Kalman filter estimates the cable’s dynamic state under stochastic disturbances, then a statistical hypothesis test on the filter’s innovation sequence flags anomalies. Unlike black-box alarms, the residual signal is interpretable and thresholds can be linked directly to false-alarm probabilities. Numerical simulations demonstrate reliable detection of localized mechanical events while ignoring ambient vibration—a key requirement for undersea environments.