Tackling global electronic warfare threats with AI

Radars and radiofrequency detectors are vital to monitor signals and track risks. But with the electronic warfare (EW) spectrum becoming more congested, it’s increasingly difficult to process and identify potential threats. Add to this the advancement of radar technology, with signal agility and concealment making it even harder to keep up. Leonardo, one of the biggest defence companies in Europe, wanted to explore whether ML can succeed where traditional methods fall short.

Our solution sought to explore and develop methods for identifying the source of a signal, especially those rarely or never seen before. Traditional methods can be complex, and often rely on significant human intervention. They can be slow to adapt to new environments and methods adversaries use for concealment. Taking inspiration from advancements in domains like audio, we developed novel ML models to distinguish signals from noise and provide meaningful output on threat levels.

This offers a significant boost in performance compared to previous methods, and was highly accurate at identifying threats quickly after encountering a new signal. The solution is superior to currently used methods and robust to potential concealment efforts by adversaries.