• Research
  • International

6G… Researchers from Bordeaux answer the call

Updated on:

While 5G is still being rolled out, electronics researchers in Bordeaux are already anticipating the arrival of 6G. By combining mathematics and artificial intelligence, they are designing electronic chips capable of paving the way for the ultra-fast communications of tomorrow. This challenge is at the heart of a European project called HERMES.

Photo : François Rivet, with a wafer of electronic chips between his hands, is the coordinator of the European HERMES project dedicated to 6G © Gautier Dufau
François Rivet, with a wafer of electronic chips between his hands, is the coordinator of the European HERMES project dedicated to 6G © Gautier Dufau

Hello?! Is your mobile phone connected to 5G? No, not yet? Because 6G is already on the horizon. Indeed, "when one generation of mobile telephony becomes accessible to all, the future one is already in the pipeline," says François Rivet, electronics specialist at the Laboratory for the Integration from Material to System (IMS)1. The lecturer-researcher from Bordeaux INP has already been working on this future for several years within the framework of the HERMES2 project led by the University of Bordeaux and funded by the European Innovation Council (EIC). Together with his fellow lecturer-researchers from Bordeaux INP, Yann Deval, Nathalie Deltimple and Éric Kerhervé, as well as Hervé Lapuyade from the university, he "helps to come up with solutions for the microchips in mobile phones. We are trying to invent the architectures of the future". You may even have a chip in your phone inspired by the work of the IMS! But why 6G when 5G is still being rolled out? Quite simply because our digital usage is exploding. "The volume of communication exchanges is constantly and exponentially increasing. Everyone wants to share more and more, images but especially video." In 2024, it is estimated that more than 2 trillion photos were exchanged in wireless communication3. The capacities of the network in 6G will be 10 to 100 times more efficient, in terms of speed, volume of data exchanged and latency, explains the scientist.

An ocean of waves: understanding the communications spectrum

Beyond sending images and videos or playing video games online, it is the cooperation between autonomous vehicles or between robots without human intervention or even the appearance of a hologram of your correspondent from your smartphone that could be made possible. Science fiction has become reality!

How can we achieve this? First of all, imagine that we live in an invisible ocean of electromagnetic waves, where each type of wave navigates according to its own rules. In the past, long radio waves of 150 to 300 kilohertz - remember the order of magnitude - advanced slowly but covered vast distances, like tall ships transporting little data. Then, FM radio (87.5 to 108 megahertz) and analogue television marked a new stage, with shorter, more dynamic waves, comparable to ships capable of carrying more information, with clearer sound and better image quality. Today, our phones operate with even shorter waves, like high-performance speedboats. 4G (700 MHz to 2.6 gigahertz) and 5G (up to 28 gigahertz) enable spectacular speeds, but require a large network of antennas to compensate for their short range.

François Rivet est spécialiste d'électronique au sein de laboratoire de l'Intégration du matériau au système (IMS) © Gautier Dufau

Communication electronics had to be reinvented.  As a researcher, it's a real intellectual challenge that means you never get bored.

François Rivet, coordinator of the HERMES project

"This spectrum of radio frequencies from 3kHz to 300 GHz is a resource that we all use, but with certain rules to ensure that we hear and understand each other," explains François Rivet. There are international agencies that regulate and supervise this spectrum. Like a maritime police force, the International Telecommunication Union (ITU) allocates frequency bands for different uses: radio, television, telephony, Wi-Fi, Bluetooth, even car radar...

With 6G, we are entering the era of supersonic jets: ultra-short waves exceeding 100 GHz. They will instantly transport a large amount of information, but will be even more sensitive to obstacles, once again requiring a dense infrastructure of small antennas. This last mile, the communication of the last few metres between the telephone and the antenna, is precisely the point on which researchers at the IMS are working in this research project. They specialise in the electronic chips that manage the radio signals of each connected object, known as RFIC (Radio Frequency Integrated Circuits). A few years ago, the Holy Grail was to create a single chip that could manage all radio signals such as Wi-Fi, Bluetooth and telephony.

Powerful yet energy-efficient microchips

Today, specific chips are being designed for each network, emphasises François Rivet. With 6G, the aim is to design radio frequency circuits that are as compact as today's (a few millimetres), while increasing their performance and reducing their energy consumption. "If we imagine that chips will be present in 10 billion devices, and that we manage to divide their consumption by 10, this could have a real environmental impact worldwide." Over the past thirty years, a race towards miniaturisation has taken place in the world of electronics to put as many components as possible in as little space as possible.

Today's figures were unimaginable yesterday: it is possible to integrate up to 1 billion transistors (an electronic component) in a chip measuring a few square millimetres. "It's so huge," says the researcher, "that the computing power of this chip is comparable to that of a mouse's brain." But it's difficult to add more. To move from one generation of telephony to another, some have assembled chips together to increase performance, but this increases energy consumption.

The challenge is therefore to create a more efficient chip, using more spectrum (100 gigahertz for 6G), that is reliable (an autonomous vehicle leaves little room for error), always compact, that consumes less and is also cheaper. "Communication electronics had to be reinvented," the scientist observes, "and it's a real intellectual challenge that means you never get bored as a researcher."

Integrated circuit connected to the outside world by gold wires (wire bonding) © IMS.
Integrated circuit connected to the outside world by gold wires (wire bonding) © IMS.

To get to 5G and beyond, the IMS researchers first changed the electronics paradigm by designing a chip that went from a temporal dimension to a frequential dimension. When you send any kind of information (image, sound, video, etc.) via a mobile phone, it is digitally recorded (encoded as 0s and 1s) on your device. But it cannot travel directly through the air; it must be converted via an electronic chip into a form, known as analogue, that is compatible with radio waves, then picked up by the receiving chip of the other telephone and converted back so that it can be seen, read, heard, etc... by the person you are speaking to. A kind of morse code, in other words! Going from the temporal to the frequential is like reading a text in its entirety versus extracting and retaining only the key words. As there are fewer words (or pieces of information) to be converted into analogue, less energy is consumed. Easy, right? It was mathematics developed by Joseph Walsh (in the 1970s) that enabled the scientists in Bordeaux to meet this challenge. And as an aside, François Rivet explains that he used the transformation known as Walsh from the American researcher, who himself was inspired by the work of another French mathematician, Jacques Hadamard, a lecturer at the University of Bordeaux between 1893 and 1897. The Bordeaux loop is complete!

A European project for disruptive innovation

But we need to go even further for 6G, and the IMS researcher had the intuition as early as 2020 that AI could help us. Taking advantage of the time made available by the COVID lockdowns, he was able to sit down and write a research project as part of the European Innovation Council (EIC) FET Open call for projects. This programme supports high-risk, high-potential research into disruptive technologies. The Bordeaux scientist collaborated with French researchers from the CEA and teams in Belgium, Greece, Lithuania and Austria to set up this European project. "I really benefited from the assistance provided by the University of Bordeaux's Proposals & Projects Support Office (SMSP) in the preparation of the application. We only had a 2% chance of success but we succeeded with funding of 3.5 million euros!"

The HERMES2 project for "radio frequency modulator and demodulator for 6G wireless systems" therefore began on September 1st 2021. The modulators and demodulators are integrated into electronic chips and constitute the modem that enables wireless communication. One of the challenges of the project was to integrate the radio frequency integrated circuits (RFIC) into a single CMOS technology chip, which is inexpensive and suitable for the mass market.

This technology has been mastered since the 1990s, but it has performance limitations for very high frequencies (such as 6G). It's a bit like trying to drive an F1 car on a dirt road. The challenge for researchers is therefore to optimise it so that it can keep up with the pace imposed by these new generations of communications, in particular by optimising the RF Front-End, i.e. all the circuits that manage radio frequency signals. When a signal - information - is amplified or converted, the output should be a perfect copy of the input, just with higher power. But in reality, the electronic components introduce distortions that deform the signal. If we go back to the example above of the text and keywords to be sent, it is as if the CMOS technology were distorting them, explains François Rivet. You can raed them here but the lettres are not in the rihgt oredr. There are systems to clean this up but they are very energy-intensive, unlike artificial intelligence which, like our brain in the previous sentence, will be able to compensate for errors and apply corrections. In the European project, AI also makes it possible to optimise the spectrum. Let's also take the example of the supersonic jets of 6G cruising on the ocean of waves. AI will act as a controller that optimises sea and air routes in real time.

Searching for gold

For example, if a 6G antenna detects that an urban area is saturated with video communication, AI can automatically adapt the frequencies and send the flows on less congested bands, just as a controller would open up express lanes for priority planes or ships.

Each country has contributed its expertise to HERMES. France has led the research into radio frequency circuits, while Austria and Lithuania are developing high-frequency modulators and demodulators. Belgium is integrating AI to optimise the performance and energy efficiency of the systems, and Greece is demonstrating it for the monitoring of radio communications in the context of maritime transport.

This very thin sheet of semiconductor material, called a wafer, can contain tens of thousands of chips. © Gautier Dufau
This very thin sheet of semiconductor material, called a wafer, can contain tens of thousands of chips. © Gautier Dufau
A wafer in measuring mode under tips, i.e. researchers are testing the performance of the circuits before they are cut out and encapsulated. © IMS
A wafer in measuring mode under tips, i.e. researchers are testing the performance of the circuits before they are cut out and encapsulated. © IMS

The circuits were planned and designed at the IMS by the researchers themselves. ‘"AI is not yet able to replace the designer in the art of goldsmithing," François Rivet explains with a smile. The circuit plans are then sent to the Franco-Italian company STMicroelectronics, the only purely European manufacturer of electronic circuits. The IMS also has a joint laboratory with the company, which facilitates fruitful exchanges for research and development. STMicroelectronics manufactures the chips from these plans on a wafer, a very thin slice or plate of semiconductor material that can contain tens of thousands of chips.

One sensitive question remains on the subject: what about the waves emitted by 6G antennas? "We are constantly surrounded by electromagnetic waves, whether from sunlight, radio, Wi-Fi or mobile phones. As with previous generations, 6G will comply with the strict standards defined by the French and European health authorities," the lecturer-researcher assures us. Impact studies will be carried out before any implementation of the future network.

In the meantime, the Bordeaux team has successfully conducted the first tests on its chips, confirming that their innovations enable the expected performance to be achieved. "Everything we had planned is working. Our results even exceed our expectations and we can now say that the project is a success!" François Rivet is delighted, while conceding that he takes a completely disconnected holiday every summer. Before returning to the laboratory, and no doubt already thinking about 7G!

1Bordeaux INP, CNRS and University of Bordeaux unit
2HERMES : High-frequEncy Radio Modulator and demodulator for 6G wirEless Systems
3sources : photutorial website

Contacts

  • François Rivet

    HERMES project coordinator
    Bordeaux INP lecturer-researcher
    Laboratory for the Integration from Material to System (IMS)

    francois.rivet%40ims-bordeaux.fr

  • Delphine Charles

    Scientific Communication Officer

    delphine.charles%40u-bordeaux.fr

  • Sophie Serhani

    Press Officer

    sophie.serhani%40u-bordeaux.fr