UAV-enabled Data Collection For Optimum Age Of Information

With the vigorous development of 5G and Internet of Things(Io T)technologies,UAVassisted mobile communication has become one of the applications with the most market development potential in the field of wireless communication.UAVs have been widely used in production and life in areas such as video shooting,environmental monitoring,3D modeling,mapping,data collection,and information broadcasting due to their excellent flexible deployment,high mobility,and low cost.In terms of data collection,how to efficiently use the limited energy carried by the UAV,reasonably plan the trajectory of UAV and optimize the data transmission strategy of UAV are the key to improving the timeliness of data and ensuring the freshness of data.This article is oriented to highly time-sensitive perception application scenarios,with the goal of ensuring data freshness,and research on the problem of UAVassisted data collection in depth.The innovations are as follows:(1)Aiming at the application scenarios of Internet of Things perception,this paper proposes a method for collecting UAV data with the best information age.First,based on the Affinity Propagation(AP)clustering method,find the best hovering point for UAV data collection,and establish the association between the hovering point of the UAV data collection and the ground perception node;on this basis,based on dynamic programming or genetic algorithm,optimize UAV flight trajectory minimizes the maximum and average information age of ground perception nodes,respectively.After theoretical derivation,this paper proves that:the UAV flight trajectory with the best information age is the shortest Hamilton path;The UAV flight trajectory with the best average information age is the shortest Hamilton path with the weight of the number of access nodes.The simulation results show that the method proposed in this paper can effectively reduce the maximum and average information age of nodes,and through effective clustering can balance the data transmission time and UAV flight time,so as to obtain the optimal information age.(2)In the Io T data collection scenario where the environmental information is unknown,this paper proposes an online data collection algorithm for the UAV based on Sarsa reinforcement learning(RL)and DQN deep reinforcement learning(DRL)framework.These algorithms determine the optimal flight direction of the UAV in real time by learning the Internet of Things topology and the data sampling characteristics of the nodes online to minimize the average age of the nodes and improve the energy efficiency of the UAV.By comparing the simulation results,we find that the Sarsa algorithm is suitable for smaller-scale networks and can obtain approximately optimal performance;compared with this,the DQN algorithm can improve data collection performance in the larger-scale Internet of Things,so it has more strong network adaptability and can also handle more complex sequence decision problems.