Research On Task Offloading Strategies For MEC Based Velocity-variable Vehicles In IoVs

With the rapid development of the internet of things technology and the continuous increase of vehicles on the road,a variety of delay-sensitive automotive applications have emerged.Many applications have strict delay requirements and need to consume a large amount of computing resources.However,the existing resources of vehicles are limited.The vehicle cannot adequately meet its needs.In the internet of vehicles scenario,mobile cloud computing is not the best choice,because the cloud server is far from the vehicle and the service quality of applications is easily affected by the fluctuation of the core network.In order to solve the above issues,mobile edge computing is a suitable solution.Mobile edge computing migrates remote cloud computing resources to the edge of the network.By deploying edge servers nearby the roadside units,it provides additional computing power for resource-limited vehicles,reduces end-to-end data transmission delay and relieves the computing pressure of vehicles.The existing research on task offloading strategies in the internet of vehicles is mainly aimed at the situation of constant velocity.On the real road,the velocity of vehicle is often variable.In this thesis,mobile edge computing technology is used to address the task offloading issues of velocity-variable vehicles in the internet of vehicles.A task offloading strategy to avoid inter-zone interruption of velocity-variable vehicles and a task offloading strategy for periodic resource scheduling based on velocity-variable vehicles are respectively proposed.The main research work of this thesis is as follows:In view of the uncertainty of the stay time of velocity-variable vehicles in the community,a task offloading strategy is proposed to avoid inter-zone interruption of velocity-variable vehicles.First,this strategy estimates the detention time of the vehicle in the cell based on the initial velocity at which the vehicle enters the current cell.And then this strategy allocates server computing resources and performs task offloading based on the estimated stay time and task delay constraints.Finally,when the task is completed or before the task is completed but the vehicle is about to leave the community,the result of the currently completed task is returned to the vehicle.The simulation results show that the proposed scheme has a significant advantage in terms of average unit task delay,average task completion rate and effectively prevents task interruption and task loss.Aiming at the problem of velocity frequently changes affecting the effectiveness of offloading task planning and computing resource allocation in the cell,a velocity-variable vehicle task offloading strategy based on periodic resource scheduling is proposed.The strategy determines the current unloading cycle of each vehicle according to the real-time velocity of each vehicle,and optimizes the task planning and computing resource allocation of the current cycle according to the dynamically updated offloading delay constraint and the average resource proportion of unit vehicles in the cell.The simulation results show the proposed scheme can effectively improve server resource utilization,reduce task offloading delay and promote the fairness of service quality of vehicles in the cell.