Distance perception in a driving simulator

Abstract

Abstract: : Purpose: The carossery often occludes lateral vision of a driver who still has to assess his own position efficiently without direct visual information about surrounding vehicles and road limits. The aim of this experiment was to test how a subject can control his approach towards several simulated car–targets in different driving contexts. We assume that their increasing complexity may influence driving performance according to the difficulty in perceiving distances properly. Method: The subjects’ first task consisted in placing their car at equal distance between two preceding cars separated either by 40 or 60 m (bisection). In a second task the subjects had to level their front bumper with the back bumper of the preceding car (alignement). Target–cars in front of the driver were either static or running at 40 or 60 km/h. Results: Results show a more precise distance perception when the task difficulty decreases. Both tasks were easier with a static simulation and in any other conditions the subjects underestimated distances. Subjects are better at adjusting their own position at 60 km/h than at 40 km/h and in bisection tasks the performance increases with smaller car distances. Conclusions: In conclusion the alignement tasks produce better performances than bisection tasks as a consequence of their lower complexity. However physical constraints due to the increase in velocity as well as shorter distances between vehicles also play a major role.

Keywords: perception • space and scene perception • vision and action