This study should point out valuable information about the process of hazard perception of (young) drivers, and evaluate their hazardperception skills to find out how well-marked this ability really is, and how they are thinking about their own hazard perception skills. For this, an eye-tracking system was used to record the gaze behavior of the drivers. In Addition, a system based on a data logger and sensors was developed to capture and log the relevant data from the vehicle. By tapping the vehicles CAN-Bus, the needed parameters (e.g. Position of accelerator, brake and clutch pedal, steering wheel angle) were gained through reverse engineering of the vehicles CAN stream. At the end, all captured data is synchronized and overlaid in the eye-tracking-video. The complete system of data-logger and sensors was integrated persistently in a driving school car. Due to the fact that all data was captured in real time traffic, every analyzed situation is unique and can’t be compared to another situation. Therefore, it’s not possible to make a statistical evaluation. So, the results have to be described qualitatively. According to all 50 analyzed situations of 25 drivers, 3 different types of hazard perception can be determined:

• An alarming number of situations where a driver doesn’t recognize a (prospective) hazard-ous situation at all.
• The driver is looking at the (prospective) hazard but keeps on driving at the same speed and doesn’t change his pedal positions.
• The driver recognizes the hazard and behaves in a safe way.

Summed up, only in 18 % of the analyzed situations the hazard perception skills are adequate or the same as they were taught at the driving school just a few months ago. Faced with the videos after the lesson, the tested drivers downscaled the self-assessment of their driving skills by around 25 % on average. So it can be shown that the hazard perception skills of young drivers are not at the sup-posed or teached level.

(EVU-members can download the full article)