Electromagnetic (EM) waves are the backbone of most modern communication systems, but seawater is a tough medium for them. Because seawater is highly conductive, EM waves lose energy quickly, and penetration depth depends strongly on frequency, salinity and temperature. At higher frequencies, such as VHF and UHF, signals are attenuated within just a few metres; even in the kHz range, reaching beyond tens or hundreds of metres is difficult and usually requires very large antennas and high-power amplifiers.

Underwater wireless communication is essential for monitoring and protecting critical subsea infrastructures, but the ocean is a very hostile environment for signals. Unlike in the air, underwater communication systems must deal with strong signal loss, limited bandwidth, long delays, low data rates, high error rates and significant background noise.

Data only becomes truly useful when it is presented in a way that users can understand quickly. The study therefore includes a visualisation environment that projects cleaned and smoothed vessel tracks onto a geographic background, using standard latitude–longitude coordinates. Route segments are coloured according to speed on a thermochromatic scale, making it easier to spot accelerations, slowdowns and other unusual behaviours at a glance.

Before applying the methodology to operational contexts, the team tested it on simulated ship trajectories where the “true” path was known. These simulations allowed them to tune key parameters—such as velocity limits and tolerance distances—and confirm that interpolation and prediction behaved as expected.

Advanced tracking techniques can be powerful but heavy, requiring specialist expertise and significant computing resources. In many real surveillance settings, however, operators need lighter tools that are fast, transparent and easy to explain to non-experts.

Marine traffic data is often noisy: signals are lost, positions “jump” on the map, and speeds sometimes look physically impossible. This is especially challenging in sensitive maritime zones, including areas close to critical and underwater infrastructures, where reliable information is essential for security and risk assessment.

The UnderSec consortium has successfully completed its first demonstration phase at Lavrion Port, Greece, held from 16 to 20 March 2026. This on-site activity brought together project partners to rehearse, deploy, and demonstrate the developed technologies in realistic port conditions.

To validate the method, the authors implemented an experiment with four sensors deployed along the north coast of Crete, from the Rethymno to the Lasithi districts. The Area of Interest was defined over central Crete, and multiple transmitters were activated within this region to test how well the algorithm could estimate their positions. Real sensor data, including bearing measurements and divergence, were used to feed the triangulation and centroid steps.

The study does not stop at theory: it presents a full algorithmic pipeline, from reading sensor data to delivering positions on a map. The algorithm is divided into clear stages, including reading and storing the input, computing bearing paths and quadrilaterals, filtering out unhelpful cases, reducing information to triads of sensors, and finally producing outputs that can be used for centroid-based position estimation. Each step is designed to keep the computation efficient and avoid unnecessary calculations in areas that are not relevant.

The paper explains triangulation in simple geometric terms: sensors measure angles, not distances, and their bearing lines intersect to form small polygons where a transmitter might be. When more than one transmitter is present, this becomes more complex, and traditional two-sensor approaches no longer suffice. The authors show that by using at least three sensors and carefully analysing how their bearing paths intersect, it is possible to separate the different triangulation areas and link them to individual transmitters.