Data sharing and privacy securing present extensive opportunities and challenges in vehicular network.This paper introducestrust access authentication scheme’as a mechanism to achieve real-time monitoring and promote...Data sharing and privacy securing present extensive opportunities and challenges in vehicular network.This paper introducestrust access authentication scheme’as a mechanism to achieve real-time monitoring and promote collaborative sharing for vehicles.Blockchain,which can provide secure authentication and protected privacy,is a crucial technology.However,traditional cloud computing performs poorly in supplying low-latency and fast-response services for moving vehicles.In this situation,edge computing enabled Blockchain network appeals to be a promising method,where moving vehicles can access storage or computing resource and get authenticated from Blockchain edge nodes directly.In this paper,a hierarchical architecture is proposed consist of vehicular network layer,Blockchain edge layer and Blockchain network layer.Through a authentication mechanism adopting digital signature algorithm,it achieves trusted authentication and ensures valid verification.Moreover,a caching scheme based on many-to-many matching is proposed to minimize average delivery delay of vehicles.Simulation results prove that the proposed caching scheme has a better performance than existing schemes based on central-ized model or edge caching strategy in terms of hit ratio and average delay.展开更多
目的为了解决车载边缘计算中用户服务质量低以及边缘节点资源不足的问题,方法结合车载边缘计算和停车边缘计算技术,提出“端-多边-云”协作计算卸载模型,并设计基于DRL的协作计算卸载与资源分配算法(cooperative computation offloading...目的为了解决车载边缘计算中用户服务质量低以及边缘节点资源不足的问题,方法结合车载边缘计算和停车边缘计算技术,提出“端-多边-云”协作计算卸载模型,并设计基于DRL的协作计算卸载与资源分配算法(cooperative computation offloading and resource allocation algorithm based on DRL,DRL-CCORA)。首先,将路边停放车辆的算力构建成停车边缘服务器(parking edge server,PES),联合边缘节点为车辆任务提供计算服务,减轻边缘节点的负载;其次,将计算卸载与资源分配问题转化为马尔可夫决策过程模型,综合时延、能耗和服务质量构建奖励函数,并根据任务需要的计算资源、任务的最大容忍时延以及车辆到PES的距离对计算任务进行预分类处理,缩减问题的规模;最后,利用双深度Q网络(double deep q network,DDQN)算法获得计算卸载和资源分配的最优策略。结果结果表明,相较于对比算法,所提算法的用户总服务质量提高了6.25%,任务的完成率提高了10.26%,任务计算的时延和能耗分别降低了18.8%、5.26%。结论所提算法优化了边缘节点的负载,降低了任务完成的时延和能耗,提高了用户的服务质量。展开更多
面对复杂的道路交通环境和网络条件,如何在多个边缘服务器之间选择合适的目标进行迁移以及避免频繁迁移导致的高额开销是一个难题。为此,提出一种依赖轨迹信息的服务迁移策略,该策略根据车辆的轨迹信息和可迁移的边缘服务器的位置信息...面对复杂的道路交通环境和网络条件,如何在多个边缘服务器之间选择合适的目标进行迁移以及避免频繁迁移导致的高额开销是一个难题。为此,提出一种依赖轨迹信息的服务迁移策略,该策略根据车辆的轨迹信息和可迁移的边缘服务器的位置信息来优化迁移策略,从而选择更合适的目标边缘服务器。在此基础上,设计了一种基于竞争深度Q网络(dueling deep Q-network, Dueling DQN)的算法进行快速决策。仿真实验证明了该策略的有效性,与其他策略比较的结果表明,该策略可以权衡时延和迁移开销,取得最小的系统总开销。展开更多
基金support by Research on Key Technologies of Dynamically Secure Identity Authentication and Risk Control of Power Business in the Science and Technology Project of State Grid Electric Power Company(No.5204XA19003F)National Natural Science Foundation of China(Grant No.601702048)
文摘Data sharing and privacy securing present extensive opportunities and challenges in vehicular network.This paper introducestrust access authentication scheme’as a mechanism to achieve real-time monitoring and promote collaborative sharing for vehicles.Blockchain,which can provide secure authentication and protected privacy,is a crucial technology.However,traditional cloud computing performs poorly in supplying low-latency and fast-response services for moving vehicles.In this situation,edge computing enabled Blockchain network appeals to be a promising method,where moving vehicles can access storage or computing resource and get authenticated from Blockchain edge nodes directly.In this paper,a hierarchical architecture is proposed consist of vehicular network layer,Blockchain edge layer and Blockchain network layer.Through a authentication mechanism adopting digital signature algorithm,it achieves trusted authentication and ensures valid verification.Moreover,a caching scheme based on many-to-many matching is proposed to minimize average delivery delay of vehicles.Simulation results prove that the proposed caching scheme has a better performance than existing schemes based on central-ized model or edge caching strategy in terms of hit ratio and average delay.
文摘目的为了解决车载边缘计算中用户服务质量低以及边缘节点资源不足的问题,方法结合车载边缘计算和停车边缘计算技术,提出“端-多边-云”协作计算卸载模型,并设计基于DRL的协作计算卸载与资源分配算法(cooperative computation offloading and resource allocation algorithm based on DRL,DRL-CCORA)。首先,将路边停放车辆的算力构建成停车边缘服务器(parking edge server,PES),联合边缘节点为车辆任务提供计算服务,减轻边缘节点的负载;其次,将计算卸载与资源分配问题转化为马尔可夫决策过程模型,综合时延、能耗和服务质量构建奖励函数,并根据任务需要的计算资源、任务的最大容忍时延以及车辆到PES的距离对计算任务进行预分类处理,缩减问题的规模;最后,利用双深度Q网络(double deep q network,DDQN)算法获得计算卸载和资源分配的最优策略。结果结果表明,相较于对比算法,所提算法的用户总服务质量提高了6.25%,任务的完成率提高了10.26%,任务计算的时延和能耗分别降低了18.8%、5.26%。结论所提算法优化了边缘节点的负载,降低了任务完成的时延和能耗,提高了用户的服务质量。
文摘车载边缘计算(vehicular edge computing,VEC)是一种可实现车联网低时延和高可靠性的关键技术。VEC可解决车载终端计算能力不足的问题,同时可减少车联网通信服务时延,避免路边停泊车辆自身大量空闲资源浪费。提出基于停泊车辆中继的车载任务卸载方案,通过合理部署边缘服务器和引入协作中继技术,可将车联网通信过程中的时延降低至毫秒级。仿真结果表明,与传统的车与基础设施(vehicle to infrastructure,V2I)卸载方案相比,基于停泊车辆中继的车载任务卸载方案可将通信时延降低20%。
文摘面对复杂的道路交通环境和网络条件,如何在多个边缘服务器之间选择合适的目标进行迁移以及避免频繁迁移导致的高额开销是一个难题。为此,提出一种依赖轨迹信息的服务迁移策略,该策略根据车辆的轨迹信息和可迁移的边缘服务器的位置信息来优化迁移策略,从而选择更合适的目标边缘服务器。在此基础上,设计了一种基于竞争深度Q网络(dueling deep Q-network, Dueling DQN)的算法进行快速决策。仿真实验证明了该策略的有效性,与其他策略比较的结果表明,该策略可以权衡时延和迁移开销,取得最小的系统总开销。
