Digital-twin-enabled 6G network autonomy and generative intelligence: Architecture, technologies and applications

Abstract

Background: The 6G network provides on-demand services and efficient governance, which is an important capability of the 6G network in the future. Providing more flexibility, availability, and dynamicity in services and network function enablement face many challenges in the 6G network due to the system complexity. To fulfill the requirement of the new generation of a digital and intelligent network, the 6G architecture with digital twin as a new enabler technology, is expected to evolve in the network autonomy and generative intelligence properties, which could optimize and adapt the network and services themselves to a variety of scenarios and deployments. Methods: To achieve 6G network autonomy and generative intelligence, we propose an architecture based on digital twin technology. This architecture integrates the digital twin network with the minimalist network and supports single-domain autonomous and cross-domain collaboration via centralized-distributed digital twin deployment, and defines the interactions between the digital twin and the service layer as well as the physical network. Results: Based on the analysis of the architecture, we put forward the following design principles for the 6G digital twin network architecture: decoupling, interactions, endogenous AI, distribution, heterogeneity, autonomous, and closed-loop. In addition, we get the key technologies achieving collaboration efficiently between the network autonomy domain and generative intelligence domain in the distributed digital twin system. Important application directions of the 6G digital twin network architecture are also analyzed. Conclusions: In this paper, the new digital-twin-enabled 6G architecture and 6G digital twin network architecture design principles are proposed. The key technologies and important application of digital-twin-enabled 6G architecture are provided.