Landside hazards detection by interferometric sensor over deployed telecom fibers for railway safety surveillance

The ubiquitous optical fiber infrastructure already installed for telecommunications purposes represents a precious asset for pervasive sensing.  The telecom fiber plan can be used not only for controlling the network integrity itself, but also for monitoring the road traffic, for large breaches and damages detection in civil structures, for the supervision of utilities health, and for the prompt detection of seismic and natural hazards.  Nowadays, systems based on distributed acoustic sensing (DAS) are successfully applied also in telecom links, to provide measurements with very high resolution and spatial accuracy in localization. In particular, DAS systems appear very attractive to operate as fiber seismometers for early earthquake detection. However, DAS back-scattering approach requires complex and expensive DSP and storage of a huge amount of data, not generally compatible with the budget of large-scale sensing applications. Interferometric solutions constitute a possible alternative, but usually they employ ultra-stable laser sources, characterized by laser coherence length longer than the sensing fiber to monitor.

For extensive applications of optical sensing in the pervasive fiber telecom infrastructure, sustainable in terms of cost, energy efficiency and reliability, we propose to adopt the interferometric approach, but constructing the interferometer itself directly embedded inside the multi-fiber telecom cable, where two fiber operate as interferometric arms. No stringent requirement is necessary for the laser sources and typical telecom DFB lasers are employed together with a simple detection scheme.

In this paper we show how the proposed “in-cable” sensor is used over the deployed telecom network to monitor geo-hazards. In particular, its operation to detect risks of landside affecting the safety of a railway in the north side of the Lake Iseo, Lombardia is presented. Two standard single-mode fibers of the 48-fiber telecom cable installed in a conduit under the sidewalk running alongside the railroad tracks by an Italian operator are employed to realize the sensor. Not only traffic events such as travelling trains, car passages on a railroad crossings and pedestrians crossing the track are detected, but above all the fall of rocks obstructing the tracks are identified. Suitable machine-learning allows to classify and discriminate dangerous events for the safety of the railway from nuisances caused by the noisy and hostile environment, that can be sources of annoying false alarms. Providing integral measurements, this interferometric sensor does not localize the event, but gives early warning of possible risks for the railway, detecting in real time landslides and notifying the alarm in a proactive way.  Other applications related to the monitoring of tunnels and viaducts, river embankments and shallow landslides triggered by rainfalls are also demonstrated.

This work was partially supported by the Italian Government through project PRIN 2022 SURENET and by the European Union under the Italian National Recovery and Resilience Plan (NRRP) of NextGeneration EU, partnership on “Telecommunications of the Future” (PE00000001 - program “RESTART”) in the project SENSING NET.