Research On Modeling And Optimization Technique Of Distributed Driving Simulator

Virtual Reality is a comprehensive technology that rised at the end of the 20^th century. It fused many branches of information and computer application technology, such as computer graphics, 3D real-time animation, computer interface, sensor, artificial inteligence and so on. It is a new mode for people to handle and interact with complex data in a visulized way through computer. When Virtual Reality technology is applied in vehicle simulation system to create virtual scene, 3D sound and movement, operators will immerse in the virtual driving environment, as is driving a real car. According to the virtual vision, hearing and touch feelings obtained from the driving environment, operators can conceive driving reactions, experience and study how to drive in the real world. Obviously, it has a lot of advantages to adopt Virtual Reality technology in vehicle simulation system, such as low-costs, safety, no limitations for time, climate and field, and he training period is short and will be very effective too.The thesis studies the VR modeling techniques of virtual scene and 3D sound according to the development of Visual Vehicle Simulation System. Firstly, the modeling tools is determined by system requirement, and VisualC++ and OpenSceneGraph(OSG) are adopted to construct simulation platform, and then the system development blue print is put forward on the basis of them. The thesis studies the three phases of graphics rendering pipeline; analyses bottlenecks of the pipeline, and give some methods to find which bottleneck it is in a given hardware architecture. The principles of some graphic rendering accelerating methods is introduced. The pipeline based on GPU and it's optimization methods are discussed too. Some important scene modeling and organising rules are brought forward according to real-time simulation system requirement. Scene modeling process, and issues about scene modeling that should be awared are discussed. In the modeling process, a lot of advanced modeling techniques, such as LOD, texture mapping, instance, external references are used to lessen the cost of system resources, improve system performance, and get perfect vision results. The thesis introduces 3D sound modeling thchnology, includes locatin principle of 3D sound, model of sound rolloff model according to distance, Doppler effect. The audio API used in the vehicle simulation system is discusses too. The design and implementation of audio module of the system are discussed in detail in the thesis.The VR modeling techniques discussed had been used successfully in system development, which fulfilled the "immersion", "real time" and "interactivity" requirement of system, and got effective simulation results. At the same time, 3D sound added into the graphic scene increase the immersion feel as it supplies the driver hearing sense.