Intelligent Scheduling of UAVs and Sensors for Information Age Minimization at Wireless Powered Internet of Things

Age of Information (AoI) has received much attention from researchers as the latest metric to quantify the freshness of data. It is necessary to jointly schedule Unmanned Aerial Vehicles (UAVs) and sensors to reduce the system AoI in wireless powered Internet of things. However, constraints on UAV flight time, charging time, and data collection time, as well as constraints of half-duplex hardware for sensors make it difficult to efficiently jointly schedule UAVs and sensors by traditional methodes. Thus, we design a multi-agent Deep Reinforcement Learning (DRL)-based UAV cooperative scheduling algorithm that jointly optimizes sensor charging time, UAV trajectories and sensor update scheduling with AoI as the optimization objective. Initially, we define the AoI minimization problem, portraying it as a Markov decision process. Then, we design a multi-agent DRL algorithm founded on factorizing value functions to address this issue. Finally, experiments demonstrate that the MAPLE algorithm can effectively coordinate the scheduling of UAVs and sensors.