Periodic Reporting for period 1 - 6G-TWIN (Integrating Network Digital Twinning into Future AI-based 6G Systems)
Periodo di rendicontazione: 2024-01-01 al 2025-06-30
Networks are becoming increasingly complex and distributed, requiring a large variety of technologies to operate. With 6G, which is now on the horizon for around 2030, it is essential to design, experiment and standardize new network architectures with more intelligence and automation.
In this context, the concept of Network Digital Twins (NDT) appears to be an ideal solution for testing a multitude of scenarios and architectural components before deploying them in the real world. However, to date, very few initiatives have focused on developing a reference architecture for NDT. Therefore, there is a need to take a major leap forward and propose new methods, simulation, and modelling tools around the concept of NDT and demonstrate their interest in tangible use cases. An important opening towards open communities is also needed to ensure these solutions' adoption and future exploitation.
In this context, 6G-TWIN will provide the foundation for the design, implementation and validation of an AI-native reference architecture for 6G systems that incorporates NDT as a core mechanism for the end-to-end, real-time optimisation, management and control of highly dynamic and complex network scenarios.
In this context, the concept of Network Digital Twins (NDT) appears to be an ideal solution for testing a multitude of scenarios and architectural components before deploying them in the real world. However, to date, very few initiatives have focused on developing a reference architecture for NDT. Therefore, there is a need to take a major leap forward and propose new methods, simulation, and modelling tools around the concept of NDT and demonstrate their interest in tangible use cases. An important opening towards open communities is also needed to ensure these solutions' adoption and future exploitation.
In this context, 6G-TWIN will provide the foundation for the design, implementation and validation of an AI-native reference architecture for 6G systems that incorporates NDT as a core mechanism for the end-to-end, real-time optimisation, management and control of highly dynamic and complex network scenarios.
During the first reporting period, the project built the foundations of a new way to design and operate mobile networks. The core idea is to create a digital twin of the network: a software copy that is constantly updated with real data. This allows researchers and operators to test new ideas, predict behaviour, and optimise performance without risking disruption of the live system.
The project defined a reference architecture that links physical networks, digital twin models, artificial intelligence, and simulation engines. It also developed a shared data space to allow secure exchange of information across partners. First models have been produced, and a graph-based engine has been selected to run them. A prototype simulation framework has been created that connects traffic and mobility simulators with network simulators, demonstrating the ability to reproduce complex situations.
Two practical use cases were prepared. One focuses on teleoperated driving, where vehicles are controlled remotely over the mobile network, and the other on reducing the energy consumed by radio access networks. Laboratories were equipped, tools installed, and first integration tests completed to prepare the demonstrators.
Finally, the project promoted open science by launching a public online space where datasets and scientific results are shared. By mid-term, more than ten scientific publications were already made available to the wider community.
The project defined a reference architecture that links physical networks, digital twin models, artificial intelligence, and simulation engines. It also developed a shared data space to allow secure exchange of information across partners. First models have been produced, and a graph-based engine has been selected to run them. A prototype simulation framework has been created that connects traffic and mobility simulators with network simulators, demonstrating the ability to reproduce complex situations.
Two practical use cases were prepared. One focuses on teleoperated driving, where vehicles are controlled remotely over the mobile network, and the other on reducing the energy consumed by radio access networks. Laboratories were equipped, tools installed, and first integration tests completed to prepare the demonstrators.
Finally, the project promoted open science by launching a public online space where datasets and scientific results are shared. By mid-term, more than ten scientific publications were already made available to the wider community.
The results achieved so far show that digital twins of mobile networks are technically feasible and already useful. They allow safer and faster testing of “what-if” scenarios, reduce the cost of experiments, and open the door to better optimisation of performance and energy use.
These results can have a direct impact on the telecom sector and society:
> Operators can use digital twins to improve reliability and to support advanced services such as remote driving or health applications.
> Energy-aware operation of networks supports Europe’s environmental goals and the wider transition to greener technologies.
> Researchers and innovators can build on the open models, data and tools already released by the project.
For these impacts to be fully realised, further steps are needed: larger and more realistic datasets provided by operators; large-scale demonstration in testbeds; close alignment with international standards; and continued support for exploitation and intellectual property management. The project has already identified three concrete innovations with potential for uptake.
The key expected results of 6G-TWIN include:
> Federated and AI-native network reference architecture that integrates multiple NDTs for real-time data analytics and decision-making.
> On-the-fly AI approaches for orchestrating network functions (NF) and services (NS).
> AI-based NF/NS for data analytics and/or decision-making to optimise network performance.
> Accurate, reliable, open and secured modelling and simulation framework for testing the 6G architecture.
These results can have a direct impact on the telecom sector and society:
> Operators can use digital twins to improve reliability and to support advanced services such as remote driving or health applications.
> Energy-aware operation of networks supports Europe’s environmental goals and the wider transition to greener technologies.
> Researchers and innovators can build on the open models, data and tools already released by the project.
For these impacts to be fully realised, further steps are needed: larger and more realistic datasets provided by operators; large-scale demonstration in testbeds; close alignment with international standards; and continued support for exploitation and intellectual property management. The project has already identified three concrete innovations with potential for uptake.
The key expected results of 6G-TWIN include:
> Federated and AI-native network reference architecture that integrates multiple NDTs for real-time data analytics and decision-making.
> On-the-fly AI approaches for orchestrating network functions (NF) and services (NS).
> AI-based NF/NS for data analytics and/or decision-making to optimise network performance.
> Accurate, reliable, open and secured modelling and simulation framework for testing the 6G architecture.