Highlights:
- Integration of AI-powered Digital Twins for next-generation 5G and 6G networks.
- Utilizes Long Short-Term Memory (LSTM) neural networks for predictive traffic management.
- Enhances network scalability, resilience, and real-time resource optimization.
- Research conducted by John Sengendo and Fabrizio Granelli, to be presented at IEEE GLOBECOM 2025.
TLDR:
A new AI-driven Digital Twin framework developed by John Sengendo and Fabrizio Granelli uses LSTM neural networks to forecast traffic and optimize resources in 5G and 6G networks—laying the foundation for fully autonomous, resilient, and high-performance wireless systems.
The rapid evolution of 5G and the emergence of 6G networks are transforming how data is transmitted, processed, and managed across the globe. However, traditional heuristic-based resource management techniques are proving inadequate in addressing the rising demands for agility, scalability, and precision in real-time communication systems. In response to these challenges, researchers John Sengendo ([link](https://arxiv.org/search/cs?searchtype=author&query=Sengendo,+J)) and Fabrizio Granelli ([link](https://arxiv.org/search/cs?searchtype=author&query=Granelli,+F)) have proposed an innovative AI-enabled Digital Twin (DT) framework that could redefine how network operators optimize complex infrastructures.
Digital Twins are dynamic, virtual replicas of physical network environments that allow operators to simulate and analyze operations without impacting the live system. The newly developed AI-Enabled Digital Twin Network (DTN) extends this concept by integrating Artificial Intelligence into the twin structure. Through the inclusion of a Long Short-Term Memory (LSTM) neural network, the system can predict traffic patterns and manage resource allocation proactively, reducing latency and improving service quality.
According to analytical experiments detailed in their study, the AI-enhanced DTN consistently outperforms baseline models in terms of traffic forecasting accuracy and optimization efficiency. By learning from historical network data, the model can anticipate bottlenecks, dynamically reallocate resources, and ensure resilient service delivery even under fluctuating demand. This advancement is particularly crucial for next-generation networks where ultra-low latency and massive device connectivity are fundamental. The findings underscore how embedding AI within Digital Twins could lead to self-optimizing, autonomous network ecosystems that operate with minimal human oversight.
Sengendo and Granelli’s approach not only positions AI-Enabled Digital Twins at the forefront of telecom innovation but also provides a roadmap for developing future-ready network architectures. As global operators transition to 6G, this research offers valuable insights into achieving smarter, greener, and more adaptive communication infrastructures designed to meet the unprecedented performance standards of tomorrow’s digital economy.
Source:
Source:
Original research paper: ‘AI-Enabled Digital Twins for Next-Generation Networks: Forecasting Traffic and Resource Management in 5G/6G’ by John Sengendo and Fabrizio Granelli, arXiv:2510.20796 [cs.NI], to be presented at IEEE Global Communications Conference (GLOBECOM) 2025. https://arxiv.org/abs/2510.20796