AoI-Aware Deep Reinforcement Learning Based UAV Path Planning for Defence Applications

Abstract

Next-gen wireless networks could greatly enhance capabilities by integrating Unmanned Aerial Vehicles (UAVs) for real-time data gathering, especially in defense. Current UAV methods focus on energy efficiency but neglect data freshness, crucial in military operations in challenging terrains. Our proposal introduces a pioneering approach prioritizing Age of Information (AoI) and energy efficiency in UAV-enabled IoT networks for such operations. Through strategic optimization of UAV locations using clustering and Deep Reinforcement Learning, we minimize average AoI and energy consumption. Our Proximal Policy Optimization UAV Trajectory Planning (PPO-UTP) algorithm, employing Deep Neural Networks (DNNs), ensures optimal decision-making based on collected data. © 2024 IEEE.