Computation Migration Optimization For Vehicular Edge Intelligence

The Internet of Vehicles is an important branch of the application of Internet of Things technology in the field of intelligent transportation.With the vigorous development of emerging technologies,such as autonomous driving,the Internet of Vehicles has attracted great attention from academia,industry and the government.Among them,the field and industry focus on the deployment of artificial intelligence applications for the Internet of Vehicles.A new type of vehicular edge computing paradigm,vehicular edge intelligence came into being.The vehicular edge intelligence refers to a new paradigm that provides artificial intelligence processing capabilities at the edge of the Internet of Vehicles networks.The vehicular edge intelligence extends the artificial intelligence processing power and strong computing capability of the traditional cloud computing center to the edge equipment,which are adjacent to the service objects,effectively expanding the computing power and storage capacity of the vehicle terminals,significantly reducing service delays,reducing communication energy and protecting the privacy.Vehicular edge intelligence is the core key to ensure the safe operation of future intelligent transportation systems.However,due to the limitations of the edge resources of the Internet of Vehicles,mobility and other characteristics,the large-scale promotion of vehicular edge intelligence has been significantly hindered.In this context,this paper focuses on the key theory and technology of vehicular edge intelligence.Aiming at the problems faced by vehicular edge intelligence,this article focuses on resource optimization and service collaboration,focusing on the key technologies and typical scenarios of computation migration,crowdsensing and data management of vehicular edge intelligence,and discussing key technical solutions for vehicular edge intelligence.Aiming at the resource expansion problem of vehicular edge computing,it is proposed to effectively use the opportunistic resources,the parked vehicles,cooperating with vehicular edge servers for joint edge computing,and at the same time study the optimal computation migration with delay constraints.Taking into account the quasi-dynamic characteristics of parked vehicles,the probability density function is used to evaluate the availability of parked vehicles.The Stackelberg game method is used to build the model,and at the same time solves the problem of computation migration optimization between the parked vehicles and vehicular edge servers.A decentralized iterative solution algorithm is proposed,which can approach the theoretical optimal value while protecting user privacy.According to the experimental results,this solution effectively expands the vehicular edge computing,and at the same time proves the superiority and effectiveness of the solution algorithm.Aiming at the problem of acquiring sensing data,it is proposed to use vehicles equipped with abundant sensor equipment to collect data,building a new type of crowdsensing paradigm,vehicular crowdsensing,and introducing vehicular edge computing technology for assistance.The vehicular edge computing technology can provide location-aware services and data management for vehicles with low latency.Among them,the main entities and functions in the vehicular crowdsensing are introduced,and an interactive protocol that can protect the privacy of vehicles is proposed.The Stackelberg game method is used to model the system,which solves the utility optimization problem of the requester and workers.The experimental results show that the proposed scheme is effective for realizing vehicular crowdsensing.Aiming at the natural disadvantages of centralized platforms,such as single node failure,it is proposed to use blockchain technology to build a decentralized,secure,and trusted data management platform.In the traditional vehicular crowdsensing,there is a central platform responsible for receiving task requests and forwarding them to the hired vehicles.This article has elaborate designed a special consortium blockchain,Ve Sen Chain,to replace the centralized central platform.Among them,the roadside units equipped with the edge server act as the miner nodes,and they are responsible for verifying the blocks generated on Ve Sen Chain.In addition,the system’s interactive protocol is designed using smart contracts.