Optimizing Function Chain Mapping Algorithms In5g Networks For The Obstacle Detection In Internet-of-Vehicles Applications

Vehicles are a common means of transportation in modern life.The traffic safety is of paramount importance,and is an urgent problem to be solved in our country and the world.Obstacle detection relies on the road condition video and the computer vision technology,which will become an important development direction for safe driving in the future.Due to the limitations of its own equipment,vehicles need to expand the perception range of obstacle detection through the information interaction in Internet of Vehicles(Io V).Under the network function virtualization(NFV)based Io V architecture,obstacle detection is deployed to the Io V in the form of a service function chain(SFC).This process quickly realizes service resource reservation,which requires allocating computing resources and bandwidth resources to service functions on demand.It is also known as the mapping of service function chains,and the core issues of NFV technology.There are some challenges in deploying obstacle detection in Io V.The mobility of the vehicle makes the Io V a dynamic topology and links.Obstacle detection is related to the safe driving and needs to reduce the mapping delay.Therefore,this paper will focus on the obstacle detection service in the Io V and study the mapping of service functions and links with the delay as the optimization goal.This paper designs the NFV service architecture in conjunction with the application scenarios of obstacle detection in Io V,providing operators with a new idea of service management and operation,and offering users with faster,more convenient and more efficient service deployment.In many ways,the traditional NFV architecture is not suitable for the Io V application scenario.The network function is not oriented to users and services,and the physical resources that vehicles and roadside base stations can supply are not considered in it.Highly centralized management and orchestration system can no longer meet the needs of the rapid deployment of Io V services.This paper analyzes the implementation process of obstacle detection in Io V,constructs the structure of the service function chain,integrates the network resources of multiple types of intelligent devices connected in Io V,and designs the NFV local resource manager to achieve service sinking.The NFV service architecture we designed improves the shortcomings of the traditional architecture,so that the Io V can efficiently schedule and manage network resources on service demand.This paper designs an efficient obstacle detection service function chain mapping model.The mapping problem in Io V is modeled as an integer programming model with the delay as the optimization goal.Considering the mobility of the vehicle,mathematical models of the vehicle network and the base station network are established with the network topology and resources,and the expression of the dynamic link is defined.A function chain mathematical model is established with resource requirements.Through user demand analysis,the objective function is to minimize the mapping delay.The function mapping constraints and link mapping constraints are described in resource scheduling.This paper designs a low-latency mapping heuristic algorithm based on hybrid genetic algorithm for the characteristics of the mapping model.We design the feasible path set initialization strategy,path change strategy and set management strategy.The algorithm introduces randomness and expands the search range of the solution,thus jumping out of the local optimal and finding the global optimal solution.We design simulation experiments and analyze the parameter setting scheme of the algorithm proposed by the control variable method.This paper compares the results with the two algorithms of Assist-Dijkstra and Tabu Search in terms of average mapping delay,link utilization,node utilization.The mapping delay is reduced by an average of 29.6 % and 8.6 %.The best mapping path can be obtained by the algorithm of this paper under different vehicle densities,and to better avoid local network congestion.This paper proposes a NFV service architecture for Io V applications,a low-latency mapping model for obstacle detection service function chains in Io V,and a low-latency mapping heuristic algorithm based on hybrid genetic algorithm.When a vehicle requests an obstacle detection service,it will realize the ondemand allocation of Io V network resources,deploy efficient communication paths,assist vehicles in driving,and reduce traffic accidents.