In its final year, UnderSec is bringing these strands together into a layered approach to underwater threat detection and decision support. At the outer layer, optimised networks of range‑free sensors, combined with automated analysis of ship movements, provide early awareness of unusual transmissions and manoeuvres near sensitive seabed infrastructure and ship hulls. Visual tools show where coverage is strong and where blind zones remain, while movement‑analysis maps highlight trajectories, accelerations and course changes that may deserve closer inspection.

Over the past year, UnderSec partners have assembled a complete toolbox to detect, track and understand activities around critical and underwater infrastructures. The project started by looking at how networks of long‑range, direction‑only sensors “see” an area, developing methods to measure their coverage, to visualise blind spots and to optimise where and how sensors should operate.

Many optimisation methods for wireless sensor networks are tested on large scenarios where the true best solution is unknown. In contrast, this article deliberately uses a small‑to‑medium test case where all possible sensor configurations can be enumerated, so the global optimum can be computed with a brute‑force search. This optimum then serves as a ground‑truth benchmark to rigorously evaluate the proposed genetic algorithm.

The study focuses on range‑free, angular‑limited sensors that only measure the direction of incoming signals, not their distance. This is highly relevant for UnderSec scenarios where inertial or direction‑only sensors detect transmitters entering a protected area, and localisation is achieved through the intersection of multiple directional bearings.

UnderSec relies on networks of sensors to detect and localise unknown transmitters around areas of interest, including critical and underwater infrastructures. In many cases, not all installed sensors can be active at the same time, due to energy, cost or operational limits, so it becomes crucial to decide which sensors to switch on and how to orient them to maximise coverage.

Before CMIM can be deployed operationally, its performance needs to be validated under controlled but realistic conditions. The team carried out laboratory experiments in a water basin using simulants of explosives and narcotics, designed to match the elemental ratios of real substances while remaining safe to handle. Measurements were performed in air, fresh water and salt water to understand how the system behaves in different environments.

Dumped ammunition, legacy chemical weapons and improvised devices attached to ship hulls are a growing concern for maritime safety and security. Corroding munitions on the seabed pose environmental and operational risks, while so‑called “narco‑torpedoes” fixed to hulls are used to smuggle large quantities of drugs beneath the waterline. Traditional inspection tools can detect anomalies, but they cannot easily tell what is inside a sealed object.

The UnderSec project is developing a new way to identify dangerous materials hidden under water or attached to ship hulls, without opening any containers. The Concealed Material Identification Module (CMIM) is an add‑on module for Remotely Operated Vehicles (ROVs) that uses pulsed fast neutrons to “look inside” suspicious objects and analyse their elemental composition in real time.

Optical underwater wireless communication (UWOC) uses light—typically in the blue‑green range—to transmit data through water. These systems can offer high bandwidth, low latency and good security because they rely on line‑of‑sight between transmitter and receiver, and are less prone to eavesdropping than many RF systems. They can also be compact, energy‑efficient and relatively low cost compared to acoustic and RF alternatives.

Acoustic waves remain the workhorse of underwater communication because they can travel much farther than EM or optical signals. Operating roughly between 2.5 kHz and 180 kHz, underwater acoustic communication (UWAC) can support long‑range links but at the cost of lower data rates, higher latency and complex channel conditions. Sound speed varies with temperature, salinity and depth, creating sound velocity profiles that bend and trap sound, while multipath reflections, Doppler effects and environmental noise further complicate reliable data transfer.