Virtual Traffic Environment Modeling Towards Experimental Platform For Driving Safety Control

Driving safety control experimental platform consists of hardware and software. The hardware system includes steering equipments such as steering wheel, accelerograph and brake pedal. The software system mainly concerns the virtual traffic environment, in which the user can interact with. Building a realistic traffic environment is crucial to the experimental data for driving safty control.There are two factors restricting the sense of reality of the virtual traffic environment, i.e. the 3D static traffic scene and the behavior of the autonomous vehicle. How to enhance the intelligence of the autonomous vehicle's driving behavior and its interacting ability are now attracting more and more researchers' attention.This dissertation studied several key technologies for building a virtual intelligent traffic environment and accomplished the prototype system. The main contributions of this thesis are presented as follows:The first, the project object, main functions and output of virtual intelligent traffic environment were studied. The implementation solution based on VR platform was discussed.The second, we made much more detailed studies on the construction of the virtual intelligent traffic environment and system simulation. The basic process of building a virtual traffic environment was discussed, including 3D static virtual scene modeling, description of the road network, user's vehicle establishing, traffic light establishing and the management of the virtual traffic scene. We use the CAD software to set up the 3D models which are finally imported into EON software package. Simplifing the road network, the road sessions are descriped by numbers of parameter, which are stored in the database. Establishing the user's vehicle and traffic light to versatile the traffic environment, we also create the scene management node to take charge of the customization and dymanic building of the virtual traffic environment, meanwhile enabling the data exchange between the virtual traffic environment and the outer program.The third, in order to build a realistic traffic environment, enhancing the intelligence of the intelligent autonomous vehicle's behavior, several key technologies of intelligent autonomous vehicle such as construction of motion model, real time traffic information obtaining based on collision detection, individual decision making are studied and the solutions are given. We built the motion model according to the Newton Laws. Binding a collision node and collision detect box with each intelligent autonomous vehicle enables them to obtain the traffic information; individual decision making module define characters of the virtual driver, sending command to the motion model, complete the driving process.Finally, based on the platform of Visual C++6.0(MFC), EON Studio 5.0, EON SDK and Windows XP, The prototype system of virtual intelligent traffic environment was accomplished. Creating the Intelligent Autonomous Vehicle node, Traffic Light node and the Scene Management node by the EON SDK, we set up a virtual traffic scene in the EON Studio. Using the Visual C++ 6.0, we build the main frame of the software sytem. We achieve the interactive ability through the object oriented programming, including the input/output of the data and the database connection. This system can implement several representative functions of virtual intelligent traffic environment such as customization of the traffic scene, road network description and the individual driving behavior of the intelligent autonomous vehicle. This system can be use for traffic accident analyzing, driving behavior research, road design and drive training.