一种基于车边云协同的车联网计算卸载策略

A Computing Offloading Strategy of Internet of VehiclesBased on Vehicle-Edge-Cloud Collaboration

随着智能交通的快速发展和车联网中数据流量爆炸式的增长,汽车终端请求卸载的任务对时延和带宽有了更加严苛的要求。在现有的云计算服务模式中,车辆可以访问云服务器来获得强大的计算、存储和网络资源,但缺点是通信传输时延较大,仅依靠云计算可能会导致过度的延迟。为了更加合理利用资源、减小时延、优化卸载策略,提出了一种基于粒子群优化算法的“车-边-云”协同卸载方案。首先通过接入点附近的软件定义网络(Software Define Network,SDN)控制器根据终端用户附近边缘节点、本地终端和云计算节点的计算资源和容量情况得出最优的卸载策略,充分利用本地、移动边缘计算(Mobile Edge Computing,MEC)设备、云端的计算资源,然后通过粒子群优化算法得出“车-边-云”各计算节点的卸载系数,即最优卸载策略。实验结果表明,相比于其他卸载策略,所提的卸载机制对时延优化效果明显,提高了计算资源的利用率。

With the rapid development of intelligent transportation and the explosive growth of data traffic in the Internet of Vehicles,the tasks requested by the car terminal to unload have more stringent requirements on time delay and bandwidth.In the existing cloud computing service model,vehicles can access cloud servers to obtain powerful computing,storage,and network resources,but the disadvantage is that communication transmission delays are relatively large,and relying on cloud computing alone may cause excessive delays.In order to make more reasonable use of resources,reduce time delay,and optimize the offloading strategy,this paper proposes a“vehicle-edge-cloud”collaborative offloading scheme based on particle swarm optimization(PSO)algorithm.The computing resources and capacity of edge nodes,local terminals,and cloud computing nodes obtain the optimal offloading strategy,make full use of local,mobile edge computing equipment,and cloud computing resources,and then use the PSO algorithm to obtain car-edge-cloud unloading coefficient of each computing node,or the optimal offloading strategy.The experimental results show that,compared with other offloading strategies,the proposed offloading mechanism has a significant effect on delay optimization and improves the utilization of computing resources.