Research On Vehicle Route Planning Method And Optimization Model In IoV

With the development of society,the number of vehicles in the Internet of Vehicles is increasing,and the problem of traffic congestion is becoming more and more serious.In people's daily travel time,the proportion of time spent on roads is increasing.Inefficient traffic control will result in waste of time and fuel,harmful carbon pollution emissions,road traffic accidents and other economic problems.In order to improve people's travel efficiency and improve current traffic conditions,it is necessary to design effective traffic flow control methods.Through effective path planning for vehicles traveling on the road,traffic diversion and control can be achieved.Therefore,this paper combines the social correlation between vehicles and the optimization of Internet of Vehicles computing resources and other issues,focusing on the vehicle path planning algorithm in the Internet of Vehicles to achieve the purpose of alleviating traffic congestion by controlling the vehicle's driving path.First,this article elaborated on the basic concepts and related theories of the Internet of Vehicles,and analyzed the architecture,organizational structure and application prospects of the Internet of Vehicles.At the same time,the basic theories and models of the evolutionary game used for IoV traffic flow control are expounded.This paper analyzes the social correlation between vehicles by considering historical and current driving information,and obtains the social clustering of vehicles.Combined with the relevant theory of evolutionary game,an evolutionary game model of vehicle route planning combined with social clustering of vehicles related to traffic state is established,and the convergence of the modified model is proved.Then this paper designs the SVRS algorithm by combining the vehicle's social clustering and evolutionary game model.In this algorithm,social clustering and game evolution games are used to predict and adjust the future driving route of the vehicle.Finally,we conducted simulation experiments to evaluate the proposed clustering method and traffic flow control based on route planning.The results show that the proposed algorithm can achieve high and good performance in analyzing vehicle cluster behavior and reducing traffic congestion.Next,considering the use of cloud centralized computing may cause transmission and inference delay in the Internet of Vehicles,this paper builds an edge intelligence based IoV architecture to distribute vehicle route planning tasks on the edge-side devices.First,a path planning model based on edge computing and multi-agent deep reinforcement learning is proposed.Then,the algorithm E-RLVRS is proposed by using edge intelligence architecture and reinforcement learning model.Secondly,preliminary optimization and prediction improved the device set and the interaction mode between devices which used to complete the path planning task.Then use the deep Q-learning to make the path planning decision of E-RLVRS algorithm.Simulation experiments verify the role of E-RLVRS algorithm in traffic flow control,and evaluate the performance of path planning tasks under the proposed edge intelligent architecture in terms of path decision time.