基于移动边缘计算的车联网卸载策略研究  

Research on Unloading Strategy of Internet of Vehicles Based on Mobile Edge Computing

作　　者：刘可欣 陈桂芬[1] 张文萍 LIU Ke-xin;CHEN Gui-fen;ZHANG Wen-ping(School of Electronics and Information Engineering,Changchun University of Science and Technology,Changchun 130022)

机构地区：[1]长春理工大学电子信息工程学院,长春130022

出　　处：《长春理工大学学报（自然科学版）》2021年第2期67-73,共7页Journal of Changchun University of Science and Technology(Natural Science Edition)

基　　金：吉林省科技厅项目(20190302103G X)。

摘　　要：随着车载设备的快速发展和日益增大的数据量,车联网在计算能力及通信能力方面面临着巨大的挑战。传统云计算虽然可以弥补车载设备计算资源的不足,但由于云服务器距离车辆终端较远,因此,难以满足一些对时延敏感的业务的需求,基于此问题,引入了移动边缘计算。首先,构建了基于5G的“车-边-云”协同网络架构,在该架构中融合了SDN等多种新兴技术,可以实现对车、边缘设备、云三方面资源的统一调度;其次,在此架构下建立了基于卸载时延的通信计算模型,并采用了基于改进烟花算法的计算任务卸载策略,其中,对烟花算法的改进主要是针对爆炸火花的产生方式及下一代烟花的选择方法,在改进之后,烟花可以实现在不同方向和不同维度的全方位搜索;最终,通过基于改进烟花算法的任务卸载策略,各个任务可以选择在最佳的卸载节点进行卸载,从而保证了时延最小化。仿真结果表明,在所提协同架构下,基于改进烟花算法的卸载策略可显著降低时延。当任务量为5 Mb时,所提卸载策略相比于其它卸载策略在降低时延性能上至少提高10%。With the rapid development of on-board equipment and the increasing amount of data,the Internet of vehicles faces great challenges in computing and communication capabilities.Although traditional cloud computing can make up for the shortage of computing resources of on-board equipment,cloud servers are often deployed far away from vehicle terminals,t herefore,it is difficult to meet the needs of delay-sensitive services.In response to this problem,mobile edge computing was introduced.Firstly,A 5G-based"vehicle-edge-cloud"collaborative network architecture was built,in which a variety of emerging technologies such as SDN were integrated,it can realize the unified scheduling of the resources of the car,the edge and the cloud.Secondly,a communication computing model based on delay was established,and a computing task unloading strategy based on improved fireworks algorithm was adopted.Among them,the improvement of the fireworks algorithm was mainly aimed at the generation method of the explosion spark and the selection method of the next generation fireworks.After the improvement,the fireworks can realize the comprehensive search in different directions;Finally,t hrough the task unloading strategy based on the improved fireworks algorithm,each task can choose to uninstall at the optimal uninstall node,thus ensured that the delay was minimized.The simulation results show that under the proposed collaborative architecture,the unloading strategy based on the improved fireworks algorithm can significantly reduce the delay.When the task volume is 5 Mb,the proposed offload strategy improves the latency performance by at least 10%compared to other offload strategies.

关 键 词：车联网 移动边缘计算 卸载策略 烟花算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]