A generative AI-enhanced digital twin framework for heterogeneous networks and smart environments

Digital Twins (DTs) have become a key technology for advancing network management and enhancing the capabilities of modern mobile networks. In this work, we present a DT framework designed to operate in heterogeneous environments characterized by diverse connectivity resources and smart devices. The framework gathers and integrates detailed information about the multiple connectivity options available for each device, enabling optimized resource utilization and real-time decision-making. The framework features a modular architecture that synchronizes physical devices with their digital counterparts, offering comprehensive insights into network performance and environmental conditions. Tested in a dynamic university smart campus-scale environment, our framework demonstrated linear scalability in resource consumption, ensuring its robustness for real-world deployments. Furthermore, by utilizing Generative AI (GenAI) capabilities such as Large Language Models, the framework streamlines analytics processes—ranging from descriptive to prescriptive—while reducing reliance on multiple models and improving operational efficiency. Finally, we highlight research opportunities in employing GenAI within DT systems, an area still in its infancy, to unlock their full potential in managing complex, heterogeneous networks and smart environments.