Research On 3D Road Network Modeling To Restrict Vehicle Driving For Edge Computing In Internet Of Vehicles

Studying the influence of 3D road network on vehicle trajectories is the important research direction in the field of intelligent transportation.In Internet of Vehicles,the driving trajectory of the vehicle is predicted according to the road structure where the vehicle is located during the task execution,which can provide a priori conditions for task offloading decision-making and cooperative node selection,so that the vehicle has super-computing power and is more sensitive to the road.The information of the surrounding environment is analyzed,decided,and controlled to achieve automatic driving.The existing research for edge computing in the Internet of Vehicles basically use the vehicle motion model that moves on a straight road or remains stationary during execution,while the driving trajectory of the vehicle in the real scene is constrained to 3D road.Therefore,studying the 3D road model that constrains the driving of vehicles can provide a more reasonable theoretical basis and experimental environment for the research on edge computing in the Internet of Vehicles.At the same time,the trajectory prediction for the edge computing in the Internet of Vehicles pays more attention to the changes in communication resources,computing resources,cache resources,etc.The relative position changes not only between vehicles and vehicles but also between vehicles and roads during task execution.It is a valuable and challenging task to model 3D roads networks.This paper focuses on the 3D road network modeling which constrains vehicle driving for edge computing in Internet of Vehicles.The main research work and innovations are as follows:(1)The existing research results lack the research on the 3D road structure model for edge computing in Internet of Vehicles to constrain the driving of the vehicle.In view of the above problems,this paper analyzes the influence of the curvature radius,longitudinal slope and length of the 3D road on the changes of vehicle speed and acceleration.Then establishes a general 3D road model which constrains vehicle driving for edge computing in Internet of Vehicles and classifies the structure of the 3D road according to the parameter variation range of the general model.(2)When vehicle is moving,it needs to analyze the road data to judge the structure of the road.This paper proposes a road structure identification method based on the three-dimensional features of discrete points.The method consists of two parts.The first part is the intersection node extraction algorithm based on the euclidean metric and the vector angle;The second part is the road classification algorithm based on the intersection topology relationship.The method can extract intersection node information and classify road types from unordered 3D discrete point data acquired by vehicles.The vehicle will continuously process the 3D road structure recognition.In order to reduce the number of recognition times,this paper proposes a region merging method based on redundant data.The method can use the data collected during the driving process and the identified road structure to provide data support for the subsequent road structure identification,combine the 3D roads and establish the road network in the driving area.(3)In order to verify the above proposed model and algorithm,this paper proposes a method to build a 3D road network experimental platform based on real 2D map.The 2D road is upgraded to 3D road by linear fitting and interpolation,and complex 3D overpass area map is established according to the 2D characteristics of the overpass in the real area.The experimental platform can mark the intersection nodes,and adjust the longitudinal slope characteristics of the 3D road according to the requirements of the experimenter to meet different research needs.Finally,this paper simulates the general model of road structure,and compares the simulation results with the Wangheqiao area in Beijing.The simulated road structure is consistent with the real road structure characteristics.Then,the 3D road structure recognition algorithm and the region merging algorithm are tested on the established experimental platform.The test results show that the algorithms have good compatibility.