Research And Application Of High Precision Real-time Rendering For Vehicles

The intelligent vehicle simulation system can be used in the intelligent automotive hardware in loop(HIL)simulation test.In order to achieve intelligent vehicle simulation,a high-fidelity digital simulation environment needs to be built.In this environment,visual scenes for visual simulation,physical scenes for dynamics and physical property simulation,and sensing layer scenes for sensing layer simulation such as lidar should be included at least.These three types of scene models contain data of different types and different attributes.Model expressions of various types of objects in the traffic environment in these three types of scenarios must have spatial consistency.The traditional scene modeling method is difficult to meet the needs of the rapid generation of scene models,which largely restricts the application of simulation technology in the field of intelligent car HIL testing.For this problem,this paper has carried out research work on the integrated modeling method of intelligent vehicle simulation environment.Firstly,an integrated modeling method is proposed in this paper,which is based on the characteristics and connections of the three scenarios of visual scene,physical scene and sensing layer scene.Firstly,the boundary of the physical scene and the sensing layer scene is obtained by using the equidistant line algorithm to transform the road plane line,and the area division of the scene is realized.Then,the Point Inclusion in Polygon(referred to as PNPoly)algorithm is used to determine which objects are included in the physical scene and the sensing layer scene.And next,the object and model files in the scene are matched by the mapping algorithm of the object position and the model file.Finally,the construction of three scene models is completed by organizing and managing the scene.Secondly,a method for modeling three-dimensional road model is proposed in this paper.Firstly,the three-dimensional interactive technology is used to complete the setting of the pile point,and the path shape of the road plane is realized through the building block method and the calculation formula of the road plane curve element.And then obtains the elevation data of the road through the three-dimensional interactive technology,and optimizes through the vertical curve.Then the 3D interactive technology is used to obtain the elevation data of the road,and the vertical curve is optimized to obtain the three-dimensional linear data of the road.And next,the road model is translated along the cross section by the equidistant line algorithm to obtain the complete road model.Finally,according to the constraint Delaunay triangulation theory,the splicing of road model and terrain model is completed by using the related implementation class of OpenSceneGraph(referred to as OSG)graphics engine.Thirdly,this paper also studies the parameters required by the model in the physical scene,which contains the geometric information and physical property information of the object.Firstly,the vertex array and the triangular face index array of the model are obtained by designing the vertex accessor.Then the simplified model of the object is obtained by the AABB bounding box algorithm.Finally,the Qt graphical editing interface is used to edit the physical properties,and the complete physical model data is obtained.Finally,this paper studies the organization management method of the scene organization.According to the characteristics of the visual scene,whose area is large and contains many objects,the optimized quadtree is used to manage it.According to the characteristics of the physical scene,the sensor layer scene,whose area is strip-shaped and contains few objects,the optimized bounding volume hierarchy tree is used to manage it.