This report describes a driving simulator experiment that was conducted to investigate driving behaviour in approach manoeuvres. The aim was to acquire more fundamental knowledge on this behaviour in order to apply it to the area of driver modelling. Driving at a constant speed, subjects approached another car which drove at a lower, also constant speed. Hence, they had to reduce their speed and then continue to follow the lead car. In addition to the braking instruction (hard versus normal braking), the subject's initial speed and the lead car's speed were varied experimentally. The results showed that Time-To-Collision (TTC) at the moment the subject started braking (TTC br) increased as a function of the subject's initial speed. In the hard braking condition, this was mainly caused by the relative speed increase. However, in the normal braking condition, there were effects of both the subject's initial speed and of the lead car's (absolute or relative) speed. The minimum TTC over the entire approach manoeuvre (TTC min) increased slightly as a function of the subject's initial speed in the hard braking condition; there was no effect of the lead car's speed. However, in the normal braking condition, there was an additional effect of the lead car's speed on TTC min. These findings show that it is not sufficient to model approach behaviour merely based on TTC or on TTC combined with absolute speed. Regarding the speed effect on steady-state car-following behaviour, it was found that especially for driving speeds lower than 40 km/h, drivers do not use a constant time headway criterion. A quadratic function gives a slightly better description of the following distance as a function of speed than a linear function.
Samenvatting