Research On Edge Computing For Vehicle-assisted Clou

With the development of network and communication technologies,the rise of new applications such as autonomous driving,image and speech processing,and online multimedia has led to an explosive increase in the demand for computing resources in mobile vehicles.In this context,traditional mobile vehicles have been unable to meet the high-efficiency and low-latency computing requirements of these emerging applications.The powerful computing power of mobile edge computing(MEC)servers provides a technical solution to this problem.To reduce data transmission time,MEC technology deploys computing and storage resources at the edge(user side).It is observed that parked vehicles(PVs)in cities have a large amount of idle computing resources,which can act as large-scale lightweight edge computing nodes to assist cloud servers to meet the secure and real-time computing needs of mobile vehicles.Therefore,different edge computing models are designed in this thesis,the cloud server and the parked vehicle cooperate with each other,and the idle computing resources of the connected vehicle are used to offload the computing task of the cloud server,and the validity of the model is proved theoretically and experimentally.The main work can be summarized as follows:(1)In this thesis,considering the interaction between different PVs,a vehicle edge computing model based on social benefits is constructed.This work first defines the benefit model of PV,which mainly considers the energy consumption generated by PV offloading computing tasks,the benefits brought by offloading cloud server tasks,and the social benefits from other PVs.Secondly,the benefit model of the cloud server is defined,mainly considering the benefits brought by the cloud server to complete the computing task,the energy cost and the economic compensation for PV.Finally,the validity of the model is proved theoretically through formula derivation.Moreover,this thesis designs a series of experiments to study the influence of model parameters on the model effect and verify the validity of the model.(2)Further,this thesis constructs a vehicle edge computing model considering both the social benefits of PV and the queuing service delay of cloud servers.Since computing tasks need to wait for processing at the cloud server,this part of the work considers the impact of task queuing delay on the benefits of the cloud server.In this model,cloud servers and PVs form a two-stage Stackelberg game and maximize their respective benefits in different stages.Furthermore,we prove the existence of Nash equilibrium in the model and calculate the optimal policies for cloud and PV.The experimental results prove the practicability of this model in improving the revenue of PV and the service efficiency of cloud server.Based on the theory of mobile edge computing,this thesis studies the cooperation mechanism between cloud servers and parked vehicles,realizes computing offloading in the environment of large-scale networked vehicles,improves the resource utilization of vehicles,and reduces the computing delay of cloud servers.By using this model,the amount of infrastructure construction can be reduced and the benefits of parked vehicle users can be improved.The research results can provide efficient,safe and responsive computing services for mobile vehicles.