New emerging applications demand performance requirements that exceed the capabilities of 5G networks, requiring a rapid evolution towards 6G networks. 6G is expected to offer data rates of the order of Tbps, time responses of the order of sub-milliseconds, and localization accuracies of the order of sub-centimeter, while meeting united nations’ sustainable development goals.
This calls for the deployment of paradigm shifting technologies and design methods, and the use of new frequency bands, including the deployment of reconfigurable metamaterial transceivers, the integration of communication and sensing functionalities, and the migration towards the sub-terahertz spectrum.
This will make 6G an extremely complex system to model and optimize.
Digital twins (DTs) and machine learning (ML) are two vital technologies to tackle the modeling and optimization complexity of 6G. A DT is a virtual replica of the 6G physical network. DTs are essential to model complex systems in real time, providing valuable insights into their behavior and performance, as well as for generating enormous amounts of training data.
ML provides advanced analytics and decision-making capabilities, enabling 6G communication systems to self-optimize, self-configure, and self-heal. The integration of DTs and ML offers a powerful approach for modeling, simulating, and optimizing 6G communication networks. It is expected to lead to the creation of a highly intelligent and dynamic network environment, where physical and virtual objects interact seamlessly, and where decisions are made and executed in real time.
TWIN6G is the first-of-its-kind staff exchange research and transfer-of-knowledge program whose aim is to build the world’s first open-access and open-source digital twin emulator to design 6G networks, integrating accurate physical models for emerging technologies and physics-based ML designs for dynamic and real-time network optimization.
