基于用户聚类的无人机集群任务规划策略  

User Cluster-based Unmanned Aerial Vehicle Swarm Mission Planning Strategy

作　　者：马文 丁飞[1,2] 赵芝因 王瑞 王诗怡[1,2] MA Wen;DING Fei;ZHAO Zhiyin;WANG Rui;WANG Shiyi(Jiangsu Key Laboratory of Broadband Wireless Communication and Internet of Things,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;School of Internet of Things,Nanjing University of Posts and Telecommunications,Nanjing 210003,China)

机构地区：[1]南京邮电大学江苏省宽带无线通信和物联网重点实验室,江苏南京210003 [2]南京邮电大学物联网学院,江苏南京210003

出　　处：《无线电通信技术》2024年第3期528-534,共7页Radio Communications Technology

基　　金：江苏省重点研发计划(BE2020084-1);江苏省“六大人才高峰”高层次人才培养资助项目(DZXX-008)。

摘　　要：为解决自然灾害引发的“断电、断网、断路”所导致的通信指挥难题,提出了一种无人机辅助网络系统,通过机载边缘服务器装载地面用户热门访问内容并发布应急通知消息。无人机基于地面服务需求,引入了Q-Learning强化学习算法进行轨迹规划并将整个系统分为无人机探索子系统和无人机服务子系统。无人机探索子系统针对单个无人机信号覆盖面积有限无法采集到区域内所有移动设备位置的问题,创建边界探索方式来确定无人机群的最优数量。无人机服务子系统通过分别对移动设备进行K-means、K-medoids、AGNES聚类选取最优聚类方式确定聚类中心,以聚类中心为导向进行轨迹规划从而尽最大可能为移动设备服务。仿真结果表明,所提无人机辅助系统具有设计可行性,确定了无人机群信号全覆盖的最小配置数量,同时得出不同聚类算法的无人机群适用场景。研究结果可用于评估地面用户的保障服务体验,为无人机辅助通信网络的架构设计与控制优化提供依据。In order to solve the communication command problem caused by“power outage,network outage,and circuit outage”caused by natural disasters,a unmanned aerial vehicle assisted network system is proposed,which loads popular access content of ground users through airborne edge server and issues emergency notification messages.Based on ground service requirements,the unmanned aerial vehicle introduces a Q-Learning reinforcement learning algorithm for trajectory planning and divides the entire system into a unmanned aerial vehicle exploration subsystem and a unmanned aerial vehicle service subsystem.The unmanned aerial vehicle exploration subsystem aims at the problem that signal coverage area of a single unmanned aerial vehicle is limited and it is impossible to collect the positions of all mobile devices in the area.A boundary exploration method is created to determine the optimal number of unmanned aerial vehicle groups.The unmanned aerial vehicle service subsystem determines the cluster center by performing K-means,K-medoids,and AGNES clustering on the mobile device respectively,and uses the cluster center as the guide to carry out trajectory planning so as to maximize the service life of mobile device.Simulation results show that the proposed unmanned aerial vehicle auxiliary system is feasible to design,and the minimum configuration quantity for full coverage of unmanned aerial vehicle swarm signals is determined,and the applicable scenarios of unmanned aerial vehicle swarms with different clustering algorithms are obtained.Research results can be used to evaluate the support service experience of ground users and to provide a basis for the architecture design and control optimization of unmanned aerial vehicle assisted communication networks.

关 键 词：无人机 强化学习 K-MEANS K-medoids AGNES 

分 类 号：TN919.23[电子电信—通信与信息系统]