Generation of simulation scenarios from the dressing of a situation

The driving license sanctions young drivers for their learning of driving procedures but does not address how to contextualize them in effective practices. Our work aims to provide drivers with a support for improving their situation awareness and to help drivers to minimize their risky behaviors by simulation of other people's errors. The simulator is designed in three steps, namely, the making of a driver typology based on risky behaviors, the making of all possible scenarios corresponding to a given situation, and evaluation of the driver’s risky behavior in the situations in selected scenarios that match this risky behavior. This paper complements the Brezillon & Tijus study about the drivers typology by focusing on the generation of scenarios derived from the instantiation of contextual driving elements. The scenarios generation system first chooses a situation specification according to the type of driver’s behavior to detect, second, then infers from contextual elements the range of normal behaviors for the situation, checks the driver‘s interpretation of the situation (i.e. practical behavior), and finally helps the driver to understand how errors can be corrected.