Road vehicle navigation through virtual world simulation

In this paper, an integrated virtual world simulator for road vehicles and networks is presented. The structure of the simulator is modular and object-oriented, where the virtual world is hierarchically constructed. It supports 2D/3D real-time graphic rendering of objects which can be visualized on multiple X-windows, and a direct “plug-and-play” of algorithms written in C/C++. The simulator provides a test bed for automated driving, driving assistance, collision avoidance and navigation strategies and tactics. Apart from providing a cost-effective way for the design and testing of concepts, the knowledge gained through the simulation may potentially be adopted by the vehicle automated driving system as rules when navigating in the real world. Our case study shows that the simulator is indeed a powerful tool for analysis and assessment of algorithm performance