无人机辅助车联网中基于车辆动态聚类的协作卸载策略  

Cooperative Offloading Strategy Based on Dynamic Clustering in UAV-assisted Vehicular Networks

作　　者：姜宇晴 李曦[1] 纪红[1] 张鹤立 JIANG Yuqing;LI Xi;JI Hong;ZHANG Heli(School of Information and Communication Engineering,Beijing University of Posts and Telecommunications,Beijing 100876,China)

机构地区：[1]北京邮电大学信息与通信工程学院,北京100876

出　　处：《移动通信》2022年第11期27-33,共7页Mobile Communications

基　　金：国家自然科学基金面上项目(62171061)。

摘　　要：在智能交通系统中,利用无人机基站可为车辆提供接入服务及广域覆盖。然而针对分布密集且动态性强的网联自动驾驶场景,动态节点智能交互频繁且须低时延以保证驾驶安全,如何动态部署无人机基站以满足网络的性能指标是一个关键问题。针对此问题,提出了基于车辆动态聚类的协作卸载策略,以降低车辆计算卸载时延,提升服务质量。首先,考虑车辆卸载请求及无人机基站的服务能力进行车簇的划分和匹配,并根据车辆位置的改变对车簇进行实时动态调整;然后,多无人机基站针对全域车辆的卸载请求基于遗传算法协作进行最优卸载决策及资源分配。仿真结果表明,与已有策略相比,所提策略可以有效降低计算卸载时延、提升服务质量。In the intelligent transportation system,UAV base stations can provide access service and wide coverage for vehicles.However,for the networked autonomous driving scenario with dense distribution and high dynamics,dynamic nodes need frequent intelligent interaction and low delay to ensure driving safety.Therefore,how to dynamically deploy the UAV base stations to meet network performance indicators is a key issue.To address this issue,a cooperative offloading strategy based on dynamic clustering is proposed to reduce the computation offloading delay of vehicles and improve the quality of service.Firstly,vehicles are clustered and matched according to their offloading requests and the service capability of UAV base stations,and each cluster is dynamically adjusted in real-time as the location of vehicles changes.Subsequently,multiple UAV base stations cooperatively make optimal offloading decisions and resource allocation based on the genetic algorithm for vehicle offloading requests in the whole area.Simulation results show that the proposed strategy can effectively reduce the computation offloading delay and improve the quality of service compared with existing strategies.

关 键 词：车联网 无人机基站 动态聚类 协作卸载 

分 类 号：TN929.5[电子电信—通信与信息系统]