Research On Key Technologies Of Data Collection And Computing Collaboration For The Optimal Age Of Information Guaranteed In Internet Of Vehicle

With the rapid development of modern information and communication technology,Io V(Internet of Vehicle)has achieved a full range of network connections,such as vehicle-to-vehicle,vehicle-to-person,in-vehicle network,and vehicle-road collaboration and so on.Taking advantage of the data collected by the Io V and the feedback information from the control center,the intelligent driving level,driving experience and system efficiency will be greatly improved.Especially for autonomous vehicles,path planning and safe driving mainly rely on the data collected by the Io V.The acquisition and analysis of these data need a lot of computing resources.As an alternative to cloud platform,Edge Network can meet the low delay requirements of such delay-sensitive tasks.However,communication resources,computing resources and storage resources of vehicular edge networks are limited.How to use limited resources to provide efficient computing support for Io V has become one of the research hotspots.In this thesis,based on the needs of the Internet of Vehicle,combined with the needs of the national key research and development projects(Low power consumption,low latency,massively connected industrial Internet application demonstration [project number2018YFB1802400])the branch project "Edge Computing Technology Research",the data collection strategy and task offloading strategy with guaranteed information freshness in the edge network are studied.The main contributions are described as follows:(1)Taking the contradiction between the comprehensiveness,fairness and limited communication resources of Io V data collection into account,we introduce the concept of Ao I(Age of Information),and propose a game theory-based information freshness optimization as well as data collection scheme.A novel stochastic game model is used to formulate the competitions among vehicles,taking into account the uncertainty of channel quality that affects the transmission success raito of packets generated by vehicles.Finally,based on an efficient Nash learning algorithm with greedy exploration,the optimal data collection strategie of vehicles,that can maximize long-term payoffs,is derived.Simulation results and performance analysis verify that the algorithm can optimize the age of information while obtaining the optimal collection strategy.(2)To tradeoff between the delay requirements of computing-intensive tasks and the limited system resources,the delay is reckon as a combination of transmission delay,computing delay and queuing delay.Based on the many-to-one matching theory,a two-stage computational collaboration scheme for joint optimization of information freshness and network utility is proposed.Computing tasks and service resources are matched at the first stage.According to the matching result of the first step,and then task transfer is performed at the second step if delay requirement of the tasks can not be satisfied.Simulation results and performance analysis verify that this scheme can maximize the network utility value and improve the success ratio of the computing task.