No description available.

Discourse about the real-world effects of automated vehicles has intensified over the last decade, but few observational studies have been made examining their integration in real traffic. This research is based on the dataset prepared by Beauchamp et al. in [1] where video footage from two pilot projects involving automated shuttles in Montreal and Candiac in 2019 was analyzed to compute safety indicators from road user trajectories. The study showed that automated shuttles have safer interactions with other road users compared to human drivers following the same trajectories. Yet, this may not be the only characteristic of automated shuttles. These vehicles are notoriously slow, 10 to 15 km/h slower than human-driven cars in city traffic [1], which on city streets is bound to influence other road users, in particular following cars. lt is therefore hypothesized that automated shuttles may have a traflic calming effect, slowing other motorized vehicles [2]. Slower speed and the predictability of automated shuttles, obeying the rules of the road and yielding more willingly to vulnerable road users (pedestrians and cyclists) may also have an impact on these users' behavior [3]: for example, cyclists may pass the shuttle, pedestrians may cross outside of crosswalks. The present study aims to explore the potential effects of automated shuttles, with their slower spceds and more predictable behavior, on the behavior of other road users. [from Introduction]