Laying the foundations for future 6G cellular technology is the aim for which the Hexa-X flagship project was born, in which the University of Pisa also collaborated, which tested the opportunity to use AI
What role will theartificial intelligence for 6G? It is good to ask yourself this already today, as we work on the creation of the sixth generation of wireless cellular technology that will succeed 5G. 6G networks will be able to use higher frequencies than 5G networks and provide higher capacity and much lower latency. Just for comparison: 6G will be able to reach a peak data rate of 1,000 gigabits/s with an air latency of less than 100 microseconds. In terms of network speed, comparing with 5G, the speed of 6G is expected to be one hundred times faster, offering higher reliability and wider network coverage.
All of this will translate into technological possibilities that are only theoretical today: 5G paves the way for the digitization and transformation of key industrial sectors such as transport, logistics, but also for the development of smart cities and self-driving vehicles. 6G will complete this scenario, opening up incredible opportunities. Future technology should allow us to enter entirely virtual or digital worlds. Not only that: in 2030, human intelligence will be enhanced by being closely coupled and perfectly intertwined with the network and digital technologies. The prospect of augmented humans can become real.
With advances in artificial intelligence, machines can transform data into reasoning and decisions that will help humans better understand and act in the world. As today’s domestic and industrial machines transform into swarms of multipurpose robots and drones, new human-machine mental and tactile interfaces to control them from anywhere should become an integral part of the future network.
All of this opens up perspectives that will have an impact on the daily lives of billions of people. But it will also have to rely on well-defined security and resilience requirements. In this regard, AI will be able to make a major contribution. As? The answer also comes from the Hexa-X flagship initiative, whose goal is to define the guidelines for the 6G network in Europe.
Takeaway
Today we are working on the development of 5G, but research and the main TLC and IT companies are looking at 6G with interest, aware of the leap in quality possible thanks to the decidedly superior performance offered by sixth generation wireless cellular technology Development of smart cities, digitization of industry are just some of the possible scenarios opened up with 5G that 6G will be able to complete, opening up the prospect of increasingly integrating physical and virtual, human and digital intelligence. To make all this possible, the European flagship project Hexa-X was born, to which 25 organizations collaborated. Among these, the University of Pisa has successfully experimented with the combination of federated learning and explainable AI
The Hexa-X lighthouse project
Hexa-X is the flagship project of the European Union for the development of the 6G network, which started in 2021 and ended (at least the first part) in these days. It was created to lay the foundations for the sixth generation of the mobile telephony network, develop the enabling technologies for the network, bearing in mind some fundamental concepts such as sustainability and transparency. Furthermore, he also paid attention to energy efficiency, but also to building trust in the use of new technologies.
Led by Nokia and Ericsson, it brought together 25 organizations, including companies and research institutions, from nine countries in a consortium. Among these is the activity of the research team of the Department of Information Engineering of the University of Pisa. Together with Intel and TIM, it has created a testbed that emulates a network beyond 5G integrated by innovative and reliable artificial intelligence systems, combining federated learning techniques and explainable AI.
Artificial intelligence for 6G, connecting physical and digital
The near future will open up new opportunities, but will also have to face increasingly demanding urgencies. The climate crisis, of course, but also possible new pandemics, social inequalities and the digital divide, geopolitical conflicts.
The advent of key technologies such as artificial intelligence for 6G, which promises a theoretical maximum speed of 1 Tbps, network virtualization and unbundling concepts can add important capabilities and design dimensions to wireless networks. An early initiation of a technology and concept evaluation, even if some of these technologies are still in a low TRL development stage, is required to understand the potential performance and impact on the overall system architecture. It is crucial to apply these new technologies to excel in new domains of use and to make them indispensable parts of the future society. As the actors of the Hexa-X consortium write:
“It is of the utmost importance to mitigate these devastating challenges, while creating opportunities for innovation-led growth and jobs. Wireless networks, central components of a digitalised society, must reflect such complex needs and opportunities and proactively deliver sustainable digital solutions, to help achieve the goals for sustainable development.
In light of these goals and challenges, the 6G Hexa-X flagship initiative was born. Its aim has been to bring together key industry stakeholders, including network vendors, operators, verticals and technology providers, as well as some of the most prestigious European research institutes and universities in the sector, creating a useful critical mass to work synergistically on research and development activities towards 6G.
Hexa-X’s vision “is to connect the human, physical and digital worlds with a fabric of key 6G enablers”. To do this, it is necessary to work on setting up the necessary infrastructures so that they can guarantee adequate guarantees. In this sense, the role of artificial intelligence for 6G becomes essential.
Towards 6G
Thus we arrive at the work coordinated by the team coordinated by Giovanni Nardini, professor of computer engineering at the University of Pisa, which has combined the potential of artificial intelligence with 6G. During the presentation of the results of the Hexa-X project at the European Conference on Networks and Communications (EuCNC 2023), the group presented the testbed and used the Simu5G network simulator, the first real-time active Open-Source network simulator , “to show that the integration of Federated Learning techniques of explicable AI models within future mobile networks is a promising approach towards high-quality and reliable innovative services to users while maintaining user privacy”, he illustrated Nardini.
If 4G has allowed everyone to communicate and to have access to internet and multimedia services above all in a practically unlimited way at any time and from anywhere, 5G is opening up opportunities such as augmented and virtual reality, in particular paving the way for digitalisation of various industrial sectors and enabling smart cities to develop.
«With 6G the goal is to continue on this path but to no longer have the vision exclusively aimed at objects, but to immerse the human being within the processes themselves, thus achieving a much closer integration between human and virtual , physical and digital. In this context, artificial intelligence plays a fundamental role».
The work of the research team of the University of Pisa
The combined use of federated learning and explainable AI and the result obtained from the “test bed” fit into this context.
«The testbed demonstrated that the integration of FL and XAI techniques, included in our research activity within the project, is not only possible but offers effective advantages», explains the DII professor himself.
In this specific case, the test activity focused on representing a realistic case of remote driving, in which a vehicle sends a video stream in real time to a remote server.
«The quality of the video sent and received by those who have to drive the car remotely must be of excellent quality, otherwise the risks of insecurity are very high».
The use case was opened to different users connected to the same network in order to share multiple experiences, training a service quality prediction model. Hence the use and confirmation of the validity of federated learning: by bringing together the training experiences of different users who use the service, it was possible to improve forecasting performance. The testbed has shown that this approach is feasible.
Artificial Intelligence for 6G – The Hexa-X Project – Giovanni Nardini (left) and Pietro Ducange, professors of the University of Pisa during the testbed presented at EuCNC 2023
But how do FL and XAI combine?
«If the aim was to bring together the experiences of different users, the problem that could arise was linked to the fact that user data, generated in real time, could be sent to a centralized body that would use them to create a model of artificial intelligence. This approach does not preserve the privacy of user data. The federated approach, on the other hand, allows individuals to create their own AI model locally, sharing it. Subsequently, an aggregator takes care of putting the information together».
Federated learning can be used with any type of AI model, but using it in conjunction with an explainable AI model adds insight to the model’s output.
The role of artificial intelligence for 6G – Demo created by the University of Pisa, Intel and TIM as part of the Hexa-X project
The future of the network beyond 5G: the role of artificial intelligence for 6G
From the point of view of the Hexa-X flagship initiative, artificial intelligence for 6G takes on a fundamental value: “through AI it is possible to control and orchestrate the performance of the network itself, which will be increasingly autonomous and capable of self-organizing and to self-optimize. In the specific case of our approach of using federated learning combined with explainable AI to predict user mobility and allow better management of road traffic and public order. Sharing data on how users move, ensuring their privacy, will allow you to understand traffic trends and manage them wisely. But it is only one approach and we are only at the beginning. «As regards the tool developed by our team, the well-founded hope is that it can serve as a tool to develop new and increasingly innovative services», concludes Nardini.
breaking latest news © (Article protected by copyright)