In one of the Advanced Crash Avoidance Technology (ACAT) projects, a computational simulation approach has been used to assess the potential benefit of three advanced Driver Assistance Technologies in a lane departure scenario. The main advantage of a computational simulation approach to driver assistance technologies evaluation is that a wide range of conditions can be explored at a comparatively low cost. Also, though multiple data sources related to traffic safety are available, few approaches make systematic and integrated use of them. Using them to validate simulation components provides a way of integrating data from various sources into a reusable format.

When using simulation, the properties of each simulated component need validation. The objective of this paper is to describe data requirements for component validation, as well as how data which meet the requirements has been identified and extracted. The basic approach of the project is to look at each simulated component and determine which of its properties influence scenario outcome. Data sources which provide input on those properties are identified, and data from them is extracted and prepared for use in the simulation. To achieve a high level of detail and accuracy for all components, data from multiple sources are used including crash databases, field operational tests, testing on test-tracks and driving simulator experiments.

The research conducted in this project shows that sufficient data can be obtained to validate the properties of the simulation components. There are limitations in available data for some sources which raises questions of representativity, but these can in principle be overcome by extended data collection. The research also shows that while extensive effort may have to go into validation the first time a simulation is developed, similar subsequent projects will require much less validation effort since the simulation components can be reused. INTRODUCTION