In a previous article I wrote about the added value of software on top of radar technology. Forendox is focused on technological security, a broad and important field, now more than ever. And although this topic is current, while also being somewhat further away from our core business, it continues to personally intrigue me.
It is therefore important to provide a complete picture and not only briefly zoom in on radar alone, as I did last time. Because, as already discussed in the previous article, there is also acoustic detection. This form of detection plays a key role today in an active conflict, and that deserves its own attention.
The story behind it is remarkable, as is the outcome. This acoustic detection emerged from a collaboration between a small number of people, including an IT entrepreneur with aspirations in the field of audio, who managed to successfully implement the technology in a completely different domain out of necessity.
As with many technological challenges, multiple factors play a role. One of them is cost. In any challenge it is important to keep costs proportional, and this case is no exception. The challenge was to detect drones over a large area, directly and accurately, in a way that requires as little cost and effort as possible, especially given the scale.
Here too, technology was used, in this case acoustic technology. A network of ‘microphones’ that sends signals back to a location where they can be processed plays a crucial role. The cost of this technique is relatively low compared to the major advantages it provides, and countermeasures against it are virtually impossible. That is what makes this technique so valuable. It had already been used before, from the coast of a European island more than eighty years ago.
The processing of these signals is fully automated. And this is where a layer of artificial intelligence comes into play today. This is not about ChatGPT, but about the foundation of AI that, in my view, represents the real breakthrough of recent years: so-called ‘agents’. These are pieces of software that have a specific task and ‘think’ about it in an almost human-like way. In this case, it is specifically about recognising drone sounds among all the other sounds a microphone picks up.
It is even known that the opposing side has tried to change the sound of their drones for this reason. That attempt was not very successful, and that is largely due to the AI agent that was used. The change only caused a small, short-lived dip in detection, which was almost immediately resolved by retraining the AI on this deviation.
In every form of defence, detection is the first step, and detection is essential. It is often even underestimated. Without detection, all subsequent steps are unthinkable. A high-tech network of sensors is therefore indispensable. The balance here is also crucial: the system must remain simple and affordable relative to the challenge so that it does not overreach its purpose. At the same time, it must be accurate. A network of acoustic sensors combined with AI agents has proven to be exactly that. Sky Fortress is, alongside drones using fibre optics instead of radio waves, a modern technological step in the spectrum of contemporary defence; a step from which there is no return.