Processing semantic information from procedural modelling rules for driving simulation

Abstract

Driving simulators are a safe and cheap way to evaluate automotive user interfaces, which is otherwise often not only very expensive but also dangerous and restricted by laws. But the creation of realistic simulation environments, especially for urban areas, can be very expensive too. So called procedural modelling is an approach to solve this problem by rule-based generation of models. Although there is a lot of information in these rules that could be useful for simulations, this information usually gets lost during the export process. This thesis compares different driving simulators and procedural modelling applications, presents an approach to make use of the information contained in the procedural modelling rules and shows how this information could be processed using CityEngine, a tool for procedural modelling, and OpenDS, an open-source driving simulator.