Research On Driverless Task Offloading Ang Caching Technology Based On MEC

In recent years,driverless cars have received extensive attention,but due to their limited battery life and computing power,it is difficult to meet the processing needs of some latencysensitive tasks or intensive tasks while ensuring battery life.By providing computing and storage resources for vehicles at the network edge,mobile edge computing can meet the realtime processing requirements of different tasks while reducing vehicle computing energy consumption.However,because the resources of the MEC server are also limited,in some cases,all tasks of the driverless cars cannot be processed.The cloud server has powerful computing resources and storage resources,which can be used as an effective supplement when the resources of the MEC server are insufficient.In view of the above problems,the research work of this thesis includes the following two parts:(1)A DQN-based driverless task offloading strategy is studied.First,under the vehicleedge-cloud architecture,a driverless task offloading model based on task priority(DTOMTP)is constructed,which needs to jointly optimize the vehicle computing power and task offloading decision to obtain the minimum delay and energy consumption of the system.Since the problem is a mixed integer programming problem,it is solved in two steps—the analytical solution of the optimal vehicle computing power is obtained through mathematical derivation,and then,under the condition of its numerical value being fixed,the optimal offloading strategy of the task is obtained based on the DQN algorithm.Finally,a simulation model is established by integrating tools such as SUMO,Pytorch and Python,and the performance of the DQN algorithm and other three common algorithms under the conditions of task load,MEC server computing power,and delay and energy consumption weight coefficient changes are compared,the performance include the delay,energy consumption,total cost of the task and completion rate.Experiments show that this strategy can achieve good results in driverless scenarios.(2)A joint task offloading and edge cache optimization strategy in driverless scenarios is studied.First,under the vehicle-edge-cloud architecture,a joint task offloading and edge caching optimization model(JTOECOM)is established,in which the MEC server is a multicore processor that can process multiple tasks at the same time,and its goal is to further optimize the system latency.and energy consumption.To solve this problem,a cloud-edge collaborative caching algorithm based on task size and content popularity is proposed,and the DQN algorithm is used to select the optimal offloading decision for tasks.Finally,a simulation model is established to compare the performance indicators of the proposed algorithm and other three common algorithms under the changes of task load,MEC server computing power and MEC server cache capacity.Experiments demonstrate the superiority and feasibility of this strategy.