AOI-aware dynamic RL-based HAP-assisted UAV charging for data collection in WSNS

The need for Unmanned Aerial Vehicles (UAVs) is becoming crucial in assisting wireless sensor networks (WSNs), such as collecting information data from sensor nodes, especially in challenging terrains, within their deadline requirement. The Age of Information (AoI) becomes then a key component in data collection, reflecting the freshness and importance of data which decreases with time if not collected. Moreover, UAVs are energy-limited and may need to be charged, which temporarily halts their data collection activities, leading to higher AoI. However, current AoI-aware studies often overlook the critical issue of UAVs' limited energy resources and assume that adequate charging infrastructure is readily available. This paper proposes an AoI-aware dynamic High Altitude Platform (HAP)-assisted UAV charging strategy for data collection in wireless sensor networks using Reinforcement Learning (RL). The UAVs' decision-making process is modeled and then enhanced through Q-learning, aiming at minimizing the overall AoI of sensors' data while considering the battery life of UAVs and extending their operational time. Simulation results demonstrate the effectiveness of the proposed approach under various system parameters.