Research On Joint Communication-Computation Handoff Strategy Based On Mobility Prediction In C-V2X And MEC Convergence Scenarios

Cellular-vehicle to everything(C-V2X)interconnects vehicles with other devices around them through cellular mobile communication(Uu interface)and direct connection communication(PC5 interface).Vehicles can exchange information via cellular networks to extend the range of sensory information.However,how to effectively process the increasing amount of massive data has become a major challenge limiting the development of intelligent transportation.Computation offloading reduces task completion latency by offloading computation-intensive tasks to cloud servers or other nearby available agent resources.Since the time taken for tasks to arrive at the cloud server is long and the network environment is constantly changing,traditional cloud computing is unable to meet the latency requirements of real-time tasks.By sinking edge servers to the edge of the network,multi-access edge computing(MEC)can reduce task completion time,alleviate core network congestion,and ensure user experience.However,for high-speed moving vehicles,frequent handoff between communication access modes and the associated computation migration between edge servers can lead to longer task completion times.To address this issue,this paper focuses on how to reduce the impact of vehicle mobility and dynamic changes of the network environment on task execution.MEC and C-V2 X convergence can provide enhancements to C-V2 X end-to-end communication capabilities,as well as support for C-V2 X application scenarios with auxiliary computation and data storage,thereby satisfying the communication and computation needs of computation-intensive vehicular applications.This paper focuses on the joint communication-computation handoff problem in the C-V2X/MEC convergence scenario.Network communication hand computation resources are divided into four layers and a resource graph is created to discuss the impact of performing network resource layering on offloading performance in a static environment.Then,the impact of vehicle mobility on offloading performance is studied using task completion time,and task completion success rate as performance indicators.Subsequently,a joint communication-computation handoff strategy based on mobility prediction is proposed considering the dynamic changes of network resources during the vehicle movement.The optimal communication link is determined by link prediction based on Hidden Markov Model,and then the communication and computation resources of the service nodes are considered comprehensively to optimize the joint handoff process by pre-migration technique to reduce the service downtime and ensure the service continuity.The average communication time,average computation time,average completion time,and task completion success rate are used as indicators of algorithm performance in the experimental sessions.The experimental results show:(1)in a static environment,the introduction of edge layer and network resource layering can shorten task completion time and improve task completion success rate;(2)for the impact of dynamic changes in network communication and computation resources on computation offloading during vehicle movement,the algorithm proposed in this paper outperforms the comparison algorithms in terms of average completion time,number of switchovers,and service downtime,and can minimize the task completion time and ensure the service quality.