Left turn gap acceptance in a simulator: Driving skill or driving style?

Abstract

This paper describes two driving simulator experiments on left turn gap acceptance at intersections. In the first experiment, participants (N = 24) had to turn left at 25 intersections and cross an approaching traffic stream of 25 equally-spaced vehicles, with a different combination of intervehicle gap time and approach speed at each intersection. In the second experiment (N = 20), gap time and approach speed differed also within each intersection stream. The first hypothesis that licensed drivers (who are assumed to have better perceptual skills) would more take into account approaching vehicles’ speed in their gap acceptance decisions than unlicensed drivers was moderately supported in the second experiment. Additionally, it was found that unlicensed drivers accepted significantly shorter gap times, significantly later in the stream of approaching vehicles. The second hypothesis that drivers who had higher self-reported violating behaviour (i.e., a poorer driving style) would be inclined to accept shorter gap times than drivers with a low violation score was not accepted. However, in the second experiment, drivers with higher violation scores were found to accept gaps that were earlier in the stream. In summary, results showed that both driving skill (i.e., having a driving licence) and driving style (i.e., violation scores) correlated with gap acceptance in a simulator. Some complicating factors, such as non-perceptual effects of driving skill, interactions between driving skill and driving style, and simulator fidelity, are discussed.