Field-testing radar innovations

Bij Thales Advanced Development & Studies leidt Ronny Harmanny een team dat zowel onderzoek doet naar nieuwe technologieën als helpt bij het op de markt brengen ervan. Een mooi voorbeeld is de functionaliteit voor dronedetectie en -classificatie die is toegevoegd aan de SQUIRE-radar.

At Thales Advanced Development & Studies, Ronny Harmanny is leading a team that both researches new technologies and assists in bringing them to the market. A case in point is the drone detection and classification functionality that has been added to the SQUIRE radar.

How can a radar tell the difference between drones and birds? Ronny Harmanny and his team at Thales Advanced Development & Studies jumped on the challenge and came up with a nifty solution. “With so-called micro-Doppler, we can observe moving parts on moving objects, enabling us to distinguish the rotating propellers of a drone from the flapping wings of a bird. Using advanced processing techniques, we can amplify the signals so we can see them from relatively far away – a few kilometers beyond the minimum instrumented range. We not only thought of the solution, but we also helped incorporate it into the SQUIRE, as part of which our invention is now being successfully deployed in the field.”

The SQUIRE is Thales’ man-portable medium-range ground surveillance radar. It can detect and classify moving targets on, or close to, the ground at ranges up to 48 km. “We use the SQUIRE to extensively test new techniques that have potential for our systems,” says Harmanny, who heads the group and takes an active part in the R&D work. “To assess the pros and cons of deep learning in a military application, for example, we can install a neural network and run experiments in the field – the SQUIRE is ideally suited for that, but we’re also looking at other radar products from Thales.”

Satellite

Having obtained a bachelor’s degree in electrical engineering from the Hanze University of Applied Sciences in Groningen and a master’s degree in computer science from the University of Twente, Harmanny joined Thales in 2000. “I started in Hengelo, but I quickly received an invitation from Eric Woonink to come and work at the new branch he’d just established in Delft. Moving to the west appealed to me and so I accepted the challenge.”

The satellite office was set up at a time when the industry was feeling a similar pinch for engineers. “Back in 2000, too, demand outstripped supply. By being close to Delft University of Technology, the idea was to attract graduates, with the aim to transfer some of them to Hengelo after a few years – I actually went against the grain,” recalls Harmanny, smilingly. “Another reason to set up shop there was the vicinity of strategic partners, such as the radar groups at TU Delft and TNO in The Hague, and an important customer, the Royal Netherlands Navy, which at the time was still in The Hague as well.”

As a radar design engineer, Harmanny contributed to several Thales products. “For the Flycatcher Mk2 air defense fire control radar, I wrote specifications and later also test documents and test software. For another project, I did some FPGA work. Initially, we mainly dealt with the development overflow from Hengelo, things that they didn’t get around to because of lack of time. Over the years, however, we got more and more innovation work, at the lower technology readiness levels. As a result, the group’s focus shifted to research, and so did mine.”

Close contact

For the R&D at the lower technology readiness levels, Advanced Development & Studies has a long-standing collaboration with TNO, formalized in the Dutch Radar Center of Expertise (D-RACE). “Within D-RACE, we basically lay the groundwork by cherry-picking potentially interesting research avenues from the scientific literature,” Harmanny explains. “Taking the findings, we think of experiments to assess their merit.”

Next to these lower-TRL projects, Advanced Development & Studies is doing a lot of higher-TRL work as well. This includes the group’s SQUIRE involvement but also integrated topside design, ie overcoming the electromagnetic compatibility challenges of the systems operating on the part of a ship above the waterline. “We look at how the radar, weapon and communication systems on a vessel interfere with one another and give advice on how to best position them to keep the disturbance at bay,” illustrates Harmanny. “We also identify the onboard radiation hazards to minimize crew exposure and we analyze a ship’s own radar signature to reduce the risk of detection.”

Additionally, Advanced Development & Studies looks at the patent portfolio for all Thales activities in the Netherlands, in close contact with patent attorney offices. And it’s responsible for spectrum management, a historically grown activity where it interacts with the Dutch authorities on transmit licenses for various Thales systems, both to support its own tests and to enable transmissions from Hengelo. Furthermore, the group is in close contact with the Dutch Radiocommunications Agency and similar bodies for military applications, advising them on possible future threats of other applicants occupying the spectrum being used by Dutch industry.

Deep learning

One of the promising research avenues being explored by Harmanny and his team is deep learning (DL). “Recently, we published a study on using DL techniques to enhance radar sensitivity. In this study, we show that we can identify individual persons in a group of fifty, based on how they walk past a radar. This can be an interesting stepping stone toward a radar security system that quickly recognizes people it saw earlier. However, we still need to cover some ground before such techniques are ready to be introduced in Thales products.”

In another research project, Harmanny’s group is looking into using deep learning to identify drones flying in a wind farm. “With its big metal blades, a wind turbine tends to completely overpower the radar signal of a small drone. There’s no way to distinguish the two using traditional techniques. We’ve shown that deep learning shows benefits for disentangling the signals and not only detect the drone but also determine its characteristics, like the number of rotors, from the micro-Doppler signature. This could prove to be very useful when operating radars around offshore wind farms.”

Employing deep learning to optimize radar operation is an important focal point for Harmanny. “Dynamically adapting the waveform, control or processing to allow us to anticipate more quickly to what we’re measuring,” as he puts it. Another area of interest is leveraging the power of phased-array radars. “Beam steering through the individual control of transmit and receive elements offers enormous flexibility. We’re not only looking at how to enable and enhance this flexibility but also at how to make the best use of it.” Last but not least, Harmanny mentions quantum computing. “Quantum computers make it possible to run new algorithms, also for radar processing. How, if at all, can we leverage this power?”

Freedom

At Advanced Development & Studies, according to Harmanny, he has the best of both worlds. “We’re part of a big company, with all the organizational benefits that that brings. At the same time, we operate relatively independently. In collaborations with partners like TNO but also in support of the Dutch Ministry of Defence, we can pursue a wide range of research topics. In national and international studies, we can even pre-competitively work together with companies that compete with us in some way and we get to take a peek in their kitchen – to some extent. This freedom is one of the things I really like about my job.” Including himself, Harmanny’s is a team of 11 people, mostly physicists, electrical engineers and computer scientists. “We’ve already got a substantial number of fresh faces, and I expect more will follow. There are lots of new developments that call for our expertise. We’re very well positioned to tackle these.”