Improved Q-Learning-Based Multi-Hop Routing for UAV-Assisted Communication

Designing efficient routing protocols for Uncrewed Aerial Vehicle (UAV)-assisted communication presents significant challenges due to rapidly changing topology, limited battery capacity, and dynamic network conditions.such as energy consumption, link quality, or latency but often overlook the necessity of an integrated approach considering a broader range of factors. This paper introduces the Improved Q-learning-based Multi-hop Routing (IQMR) algorithm that facilitates energy-efficient, and reliable data transmission in UAV-assisted communication. IQMR achieves this by selecting the optimal next-hop node to ensure efficient energy utilization, reliable packet delivery through collision avoidance, and adaptive network reorganization to maintain connectivity without relying on predefined UAV paths. To the best of our knowledge, IQMR is the first to employ a multi-objective framework that captures the inter-dependencies between network parameters and UAV operational states while leveraging $Q(\lambda)$ learning to make routing decisions, ensuring reliable communication in dynamic environments. Results show that IQMR demonstrates a 36.35% improvement in energy efficiency and a 32.05% increase in data throughput over existing methods.