Research On Edge Caching Algorithms For Internet Of Vehicles

In recent years,with the development of the intelligent transportation and the internet of vehicles(IoV),the amount of network data in IoV shows a trend of rapid growth,and a large number of data requests with low delay requirements need to be satisfied.At the same time,with the rapid growth of the number of intelligent vehicles,the edge caching methods in IoV has got lots of attention.Intelligent and connected vehicle(ICV)and road side units(RSU)serve as the edge cache nodes by carrying mobile edge servers,providing content delivery service,avoiding high delay caused by obtaining content from backbone network and relieving the backhaul links pressure.However,due to the high mobility of vehicles,the uncertainty of user requirements and the limited storage capacity of edge cache nodes,the cache allocation has become an important issue in the edge network environment.The main work of this paper is as follows:(1)In the simulation of the edge cache of IoV,it is important to match the vehicle trajectory points to the road network.This paper proposes a matching algorithm between the vehicle trajectory points and the road network,and constructs the spatiotemporal dynamic speed network of vehicles on the road network through a large number of trajectory data.(2)Aiming at the complexity of edge collaborative cache allocation in IoV,an edge collaborative cache algorithm based on deep reinforcement learning is proposed.This paper first introduces the collaborative cache model of ICV and RSU,then defines the cache allocation problem with Markov decision model,solves it with DQN network and finally carries out simulation experiment.Experimental results show that the proposed algorithm achieves 91.1%cache hit ratio.Compared with the traditional cache algorithm,it can effectively increase the cache hit ratio and reduce the system cost.(3)Aiming at the cache update lags of vehicles cross regions,an edge collaborative cache algorithm based on vehicle movement prediction is proposed.Through the analysis of vehicle trajectory data,the future moving direction of vehicles is predicted,and the hot spots on the road network are mined.Combined with the predicted direction of vehicle movement and the location of hot spot area,a mechanism of cache fast update in advance is provided before the vehicle enters the hot spot area.Experimental results show that,compared with the traditional algorithm,the cache hit rate of the proposed algorithm is 8%higher than that of the traditional algorithm.