Automatic Road Networks Generation Dedicated to Night-Time Driving Simulation

Driving simulation is used for more and more complex studies nowadays that are mostly linked with safety and comfort. Absence of risk, reproducibility of situations and strict control over experimental parameters: these advantages make the use of a simulator crucial to study driving in night-time conditions. The development of adaptive front-lighting system is a field where driving simulators have been lately used. Since the preparation of experimentation requires to iteratively improve the database according to the expected scenario, the process of generating road networks database has to be as easy as possible, with the constraint that the result matches the required photometric and perceptive quality. To suit night-time simulation, 3D databases must meet some strict requirements: database must be finely tessellated, all polygons must have a similar size once projected on screen, and, in order to get a satisfying refresh rate, the number of computed polygons per image must stay acceptable. Moreover, texture used in such a database must be properly set to match the particular night-time conditions. In order to satisfy all the constraints Institut National de REcherche sur les Transports et leur Securitee (INRETS) and OKTAL have set up processes to easily generate or regenerate 3D database even when modifying parameters such as: road description, transverse profiles, longitudinal profiles and crossroads. At the core of this process are algorithms based upon transverses profiles application, automatic texture application and automatic levels of detail generation and usage. These developments concern projects aiming to improve the use of innovative lighting technologies, as well as those studying, for example, the visibility of pedestrians in night conditions.