Stefan Valentin, Professor of Mobile Networks at the Darmstadt University of Applied Sciences, and his team are working on better protection of the 5th generation mobile network (5G). The scientists have in mind the network that millions of people use every day when they make calls on their mobile phones. But they are most interested in corporate networks, where 5G technology has many applications in controlling complicated, time-critical workflows – for example in large logistics centres, the automotive industry, hospitals, and in the future possibly also in public security and the fire service. So it is used in the critical infrastructure – which can, with the necessary criminal energy, be paralysed using simple jammers. Valentin’s watchdog NERO is set to change that.

By Christina Janssen, 19 November 2023

Stefan Valentin still thinks about Angela Merkel’s mobile phone sometimes. In 2013 the Chancellor declared that there should be no spying among friends, after the US National Security Agency (NSA) had tapped her mobile telephone. “Unfortunately, it’s technically simple,” says Valentin. For this reason distrust is right at the top of his list of virtues: “Don’t trust any device more complex than a toaster,” is his advice. “A toaster has an on-off switch, and when it gets hot I know that it’s working.” And a mobile phone? “It can do what it wants without you noticing.”

Professor Valentin has been teaching and researching at the Darmstadt University of Applied Sciences since 2018. His current research project focuses on – surprise! – security. Or to be more precise, on the security of 5G networks. “Right now there are no European manufacturers of 5G modems,” he explains. “So every device contains components from other continents. We cannot resolve this supply chain problem, and therefore we are providing 5G base stations with a watchdog.” Valentin and Professor Martin Stiemerling have joined forces for this project as the joint heads of the Networking Technologies research group at the h_da. Other participants in the “ADWISOR5G” project are the Padova, the Fraunhofer-Institute for Integrated Circuits (IIS) and two industrial partners – the network service provider Mugler and the equipment supplier albis-elcon. Their 5G watchdog already has a name: the Network Real-time Observer (NERO) is intended to raise the alarm when something suspicious is detected in the network. The Federal Office for Information Security is supporting the project for 18 months with over EUR 1.8 million, EUR 511,000 of that going to the h_da.

The problem is that it’s easy to paralyse wireless networks – using simple jammers, for instance, which drown out signals from the wireless network by creating radio noise. Valentin calls this the “sneaker method” – in contrast to more sophisticated jammers that block specific frequencies. “That would be the ‘stiletto’ variant. If I tread on someone’s foot with a stiletto heel, it hurts more than a soft sneaker,” Valentin explains. He has received several awards for his creative metaphorical explanations of his research projects.

No matter whether they are using stilettos or sneakers, the hackers are usually interested in sabotage or espionage. “Either I want to make life difficult for a competitor, or I use the jammer to distract from my real purpose of stealing data,” Valentin explains. Most attacks on 5G networks are therefore not directed at private smartphones. If a user loses mobile coverage, a denial-of-service attack using a jammer could be the reason. It’s easy to smuggle these small jammers into a company. They have an antenna that fits in a trouser pocket and they are quick to build. “My students can put them together in one minute,” says Valentin. “The devices can easily be included in a returned parcel – so the transmitter quickly reaches the warehouse.” Once the jammer is in place, it can go about its mischief without anyone intervening.

And this is where NERO comes into play. Professor Valentin’s team is training an AI system using unsupervised machine learning to recognise deviations from a wireless network’s normal state. To start with, watchdog NERO is “fed” with regular radio signals using graphs representing an undisturbed QAM-signal, on which modern telecommunication systems such as 5G, WiFi and DSL are based. So NERO learns what a normal radio signal looks like. But, like a real dog, NERO is also chipped. It contains a radio chip programmed by Valentin himself, which makes it vigilant. The chip is completely software-controlled – the scientific term for this is “software defined radio” (SDR) – and reports deviations from the undisturbed signal.

“NERO’s job is to sound the alarm when something’s not right.” To train and test the system, Valentin and his researchers Michael Birger and Matteo Varotto have built their own 5G network in the laboratory, installed several different jammers, and set NERO to work. It’s been a great success: in hundreds of test runs, NERO has recognised the attacks in a good 96 percent of cases. When something like this occurs in real life, the jammer has to be located and removed as quickly as possible. “Basically, that’s the only way to terminate the attack,” Valentin explains. “Our colleagues at the Fraunhofer Institute are therefore working hard on radiolocation.”

What Valentin and his team are planning is largely new. “Up to now most systems have been analysing network data. But we’re warding off attacks at the physical layer.” The great advantage is that the laws of physics are non-negotiable. When watchdog NERO is up and running, it will be impossible to trick it. So when NERO is around, there will not be an ‘arms race’ like that between hackers, who are always creating new computer viruses, and IT security experts. Later on, the project team will also tackle detecting the more complex disruptive ‘stiletto’ attacks. By the end of the project, a prototype technological watchdog should be ready for a company to develop it to market maturity.

Some firms are already interested, which pleases Valentin, himself a radio ham: “That’s the nice thing about this project – it has practical relevance. We work directly with the technology, but we’re also quickly back in the real world. It’s a nice example of applied research.” So what about Merkel’s mobile phone? Of course Stefan Valentin knows not only that it was tapped, but also exactly how it was done. “A text message with a hidden piece of code was sent to the Chancellor’s phone. The code said, ‘Call this number, but don’t let the user know that a call is being made.’ This activated a secret channel.” Valentin has only one thing to say: “NERO would have raised the alarm.”