Edge-Assisted Intelligent Routing Framework for Reliable Data Collection in UAV Networks Using Reinforcement Learning

This paper proposes an Edge-Assisted Intelligent Routing Framework (EIRF) that integrates Reinforcement Learning (RL) with edge computing, aligning with Software-Defined Networking (SDN) and Network Function Virtualization (NFV) principles for adaptive and resilient UAV communication. The predictive module for mitigating failures utilizes edge resources to predict gateway issues and proactively adjust network paths. Also, the system employs a Deep Reinforcement Learning (DRL) algorithm to dynamically allocate resources and ensure seamless transitions between gateways, maintaining reliable communication under varying network conditions. Extensive simulations demonstrate that EIRF significantly reduces switching latency, improves throughput, and enhances network resilience compared to traditional methods such as Optimized Link State Routing Protocol (OLSR) and batman-adv.