| With the development of science and technology,various new types of automotive applications emerge in an endless stream,such as AR navigation,human-computer dynamic interaction,etc.These applications have strict requirements for the delay and reliability of data forwarding,but traditional cloud computing cannot meet all new types of applications.Therefore,low-latency and highreliability data forwarding in vehicular communication faces huge challenges.With the development of mobile edge computing technology,edge caching provides a feasible solution to low-latency and high-reliability data forwarding in the Internet of Vehicles(Io V).However,in practical scenarios,the number of edge servers deployed and their storage capacity are limited,but the communication range of each single vehicle is also limited.Besides the cached content in some vehicles is not fully utilized.In this case,how to design a method that can expand the range of vehicle data forwarding in the Io V,and also an edge caching solution that can reduce the delay of data acquisition,to solve the problem of limited communication range of a single vehicle,and to meet the needs of new applications for data.The higher requirement of transmission performance has become an urgent problem to be solved.This thesis focuses on the edge network environment of the Io V,focusing on the edge caching based on vehicle clustering in the Io V.The main research work of this thesis is as follows:(1)The fast density based clustering algorithm of vehicles in the Io V is studied.The algorithm firstly creates an initial cluster based on the similarity between the vehicle and the existing cluster as the edge server receives the relevant information of the vehicle.Then,the initial rough clusters are sorted,and all nodes in the initial cluster are regarded as micro-clusters,and the cluster merging operation is performed.After the merging operation is completed,the sorting ants are created,and the micro-clusters are transported according to the similarity between clusters by using the principle of ant accumulation of corpses.When the clusters are sorted,the cluster heads are set,and cluster maintenance operations are performed during subsequent movements.The simulation results show that compared with the VANETs clustering algorithm based on ant colony optimization(CACONET)and DBSCAN clustering algorithm,this scheme reduces the number of clusters by about 30%,increases the average cluster head duration by about 28%,and increases the average cluster member duration by about 10%.(2)The edge caching scheme based on vehicle clustering is studied.This scheme aims to minimize the content acquisition delay of the requested vehicle within the communication range of the edge server,and studies the edge cache decision optimization algorithm in the Internet of Vehicles,which includes the optimization of vehicle-related variables and the optimization of cache decision variables.In the vehicle-related variable optimization problem,firstly,the vehicles within the communication range of the edge server are divided into the requesting vehicles and the cached vehicles,and secondly,the requesting vehicle decides whether to forward the content through the cluster head according to the communication range of the cached vehicle in the V2 V link.In the V2 I link,whether to forward the content through the cluster head is determined according to the communication time of the cluster head and the content acquisition delay.Finally,the vehicle is requested to select a link for transmission according to the content acquisition delay of the V2 V link and the V2 I link.On the optimization problem of caching decision variables,according to the results of the optimization problem of vehiclerelated variables,the minimization of content acquisition delay is transformed into a knapsack problem,and the dynamic programming algorithm is used to solve the problem,and finally the results of caching decision variables are obtained.The simulation results show that compared with the random cache scheme,the task acquisition delay is reduced by at least about 100 seconds,and the cache hit rate is increased by about 50%. |
