Automatic processing of a car driver eye scanning movements on a simulated highway driving context

This paper deals with the description of a system performing an automatic data extraction from the visual activity recordings of a driver in a simulated driving context. Main modules of this system are described with their related principles and problems. From the hardware point of view there is the driving simulator, the eye tracker and the developed synchronization device. From the software one, there is the numerical raw data filtering process, the ray tracing module and the decision software. The validation of this system puts forward its good capacity to reproduce the manual data extraction of an expert operator from video recordings. Its first application to a long duration driving experiment points out the large amount of time spared in data extraction thanks to this system.