Research On High-precision Network Model Of Road Traffic For Autonomous Driving

With the rapid development of intelligent transportation and artificial intelligence technology,autonomous driving has become an inevitable trend of automotive automation.It has already become a current central issue in automotive technology.The car itself cannot make autonomous driving.It depends on the Internet of Vehicles technology for collaborative control and high-precision maps to improve positioning accuracy and assist the environment recognition.Therefore,to develop autonomous driving technology,the corresponding highprecision map is essential.The high-precision network model of road traffic for autonomous driving is studied,which is the road network model in high-precision map.Firstly,the advantages and disadvantages of the existing road network models are analyzed and summarized.Considering the data content and data organization requirements of the road network model,with discussion on topological relationship expression of the road network,a high-precision network conceptual model of road traffic for autopilot is proposed,which consists of basic road network,ancillary facilities and traffic events.Upon that,the logical model design is completed.Secondly,the construction technologies of road network model is studied.From geometric data to topological data,road level to lane level to lane level,current construction method is improved for the designed model.A road-level road network construction method by combining triangulation method and unit circle rolling pursuit method is proposed.The generation and topology construction method of lanelets and lane groups based on lane separation line marks is put forward.Finally,based on the theoretical basis of the above model design and construction technology,based on.Net framework and ArcGIS component library,by using real road surface vector data,the model is constructed through program development.And through the simulation of route guidance on both road-level and lane-level,the feasibility of the model and construction techniques are verified.