Impacts of an intersection assistance system - Social interaction between drivers and motorcyclists

In a simulator experiment, car drivers and motorcyclists were navigated on collision course towards intersections. Both drove in a common virtual environment, but separate driving simulators. The penetration rate of an informing intersection assistant was varied: no assistance, driver assisted, motorcyclist assisted and both drivers assisted. In this paper, the velocity signals of both road users were used, to detect differences in their interactive behavior. These differences should be visualized and quantified using the cross recurrence quantification analysis. First, the distance to collision plots of motorcyclist and car driver were examined. This resulted in the identification of two situation groups: yielding and not yielding situations. Next, the reaction times of both drivers were calculated as a point of reference for later analysis. The reaction times between the experimental conditions (assistance level of road users) differed significantly. Subsequent evaluation of cross recurrence quantification measures also showed significant differences between the assistance configurations both for the global and the yielding perspective. Yet, only results for the yielding perspective showed the expected results. Two reasons are suspected for this: 1) test subjects braked stronger in yielding situations and 2) the situations, where the driver didn't react, are less common in this study. Two improvements for further research are proposed: a head-up display for motorcyclists to improve their awareness of the intersection assistant and a larger sample to lower variation in the yielding perspective.