Research On Caching Strategy In Vehicular Network

In recent years,with the rapid development of intelligent transportation systems,a large number of applications related to vehicle networks have appeared,the rapidly changeable topology lets vehicle network fail to provide vehicles with good communication environment.The introduction of caching technology into vehicle network can effectively reduce the user’s request delay as well as improve the success rate of user’s requests.However,the number of users in vehicle network increases rapidly,and the user’s requests also become more diversified,so how to design the caching strategy of the caching nodes to meet users’ requirements has become the key technology of vehicle network.This thesis studies the caching technology in vehicle network,and based on the prediction about vehicle’s moving trajectory realized by neural network,we propose a proactive caching strategy based on the predicted trajectory of vehicle cluster.According to users’ interest preferences,we also propose a bus caching strategy based on the content popularity in different streets.Based on the requests probability of contents in different street as well as the distance between roadside units(RSU)and streets,we also propose a RSUs’ placement and caching strategy.The main research contents and innovations of the thesis are as follows:(1)Aiming at the problem of low content acquisition success rate caused by rapid changes in vehicle network topology,a proactive caching strategy based on cluster’s trajectory prediction is proposed.The RSU divides vehicles into different logical clusters according to their predicted trajectories,actively caches the requested content in the cluster that is about to meet the requesting vehicle,the requirement of communication links’ quality is reduced by dividing content into pieces and caching them in multiple vehicles.On this basis,based on the number of vehicles in the cluster and the vehicles’ moving characteristics,this thesis proposes an optimization algorithm for caching content in the cluster,which maxmizes the successful content transmission probability,and an expression for the probability is derived.The simulation results show that the proactive caching strategy proposed in this thesis improves the success rate that users can obtain the requested content.(2)In order to solve the problem that outdoor users in urban hotspots have a long delay in obtaining requested contents,based on bus’ s fixed travel trajectories and fixed running timetables,a bus caching strategy based on the popularity of content in different areas is proposed.The neural network is used to classify contents,and according to contents’ regional popularity and the cross entropy theory,the cache ratio of content with different categories is optimized to improve the user’s request hit rate and reduce request delay.And the optimal cache ratio is obtained by the particle swarm algorithm.In addition,based on the statistical characteristics of the bus running time,the probability that the user successfully obtains the requested content is deduced.On this basis,a utility function is defined by considering the probability as well as the cost time that users successfully obtain requested contents,the sum of all users’ utility function is to be maximized by optimizing the content delivery strategy.The simulation results show that the caching strategy proposed in this thesis can reduce user’s request delay.(3)Aiming at the problem that the content transmission distance is too long,and in consideration of high cost of deploying full coverage RSUs,a RSU placement and cache strategy based on request probability of content in different streets is proposed.According to the request probability of content in different streets and the distance between streets and RSUs,the optimization problem that minimizes the content transmission distance is established,and the optimal caching strategy of RSU is deduced by solving the optimization problem with the help of the traversal method.On this basis,RSUs and vehicles make up of the multi-hop network that supports content transmission process.In order to reduce the number of contents’ forward,our thesis preferentially selectes the driving vehicle as a relay.At the same time,in order to make full use of the cache resources of the parked vehicle,our thesis lets those vehicles cache contents according to content request probability and content transmission distance.In addition,based on the statistical characteristics of the distribution of parked vehicles,the expressions of communication time between moving vehicle and parked vehicles as well as the expressions of content forwarding times during transmission process are derived.Simulation results show that the proposed caching strategy reduces deployment cost and users’ request delay.