Deep Learning Empowered Service Selection in FANETs: A 5G-Enabled UAV-Cloud Architecture

Flying Ad-Hoc Networks (FANETs), commonly referred to as drones or Unmanned Aerial Vehicles (UAVs), are increasingly transforming various industries by supporting numerous applications. In this context, a UAV can function as a cloud service provider, offering services such as memory, computing, and network connectivity to consumer UAVs. This work introduces an innovative 5G-enabled UAV-Cloud Architecture empowered by deep learning for service selection. The architecture integrates two main modules: (i) A game theory-based module, which establishes an interactive game between provider and consumer UAVs, generating datasets to guide consumer UAVs in selecting suitable providers; and (ii) A deep learning-based module, leveraging the generated datasets to predict optimal service providers. Numerical simulations validate the proposed architecture, highlighting its superior prediction accuracy, reduced time and message complexity, and minimized service discovery and consumption delays compared to existing methods. These advancements are attributed to the synergy of game theory, deep learning, and 5G communication capabilities.

• flying ad hoc networks
• service selection
• 5G
• unmanned aerial vehicles
• game theory
• deep learning
• service discovery

• 9 — Industry, innovation and infrastructure

• UAV Applications and Optimization
• Service-Oriented Architecture and Web Services
• Mobile Ad Hoc Networks
• flying ad hoc networks
• service selection
• 5G
• unmanned aerial vehicles
• game theory
• deep learning
• service discovery
• Aerospace Engineering
• Information Systems
• Computer Networks and Communications
• Engineering
• Computer Science