Understanding interactions between Automated Road Transport Systems and other road users: A video analysis

Abstract

One of the proposed future benefits of automated vehicles (AVs) is the provision of an alternative public transport solution, which would increase the number of mobility options in busy, urban environments. If these types of public AVs are to move efficiently through shared spaces, there is a need for them to interact and communicate with other road users in a comprehensible and predictable manner, and thus an understanding of the interaction requirements of other road users is needed. The current study addressed this issue by using video-analysis techniques to identify the typical interaction patterns which might arise between these public transport AVs and other road users, based on interactions observed during CityMobil2 demonstrations in La Rochelle in France and Trikala in Greece. The purpose of this analysis was to evaluate the associations between contextual factors such as environmental features and road user characteristics, and common AV interaction categories. A combination of manual and automated video coding processes were used to categorise all observed interactions between the AVs and other road users during 22 hours of video footage across the two locations. The results indicate that both environmental and road user factors had a major impact on the type of interactions that arose between AVs and other road users. Features such as the width of the road and the presence or absence of infrastructure such as zebra crossings had an impact on road users’ trajectory decisions while approaching an AV. Where possible, vulnerable road users, such as pedestrians and cyclists, appeared to leave as much space as possible between their trajectories and that of the AV. However, in situations where the infrastructure did not allow for the separation of traffic, risky behaviours were more likely to emerge with both pedestrians and cyclists travelling closely alongside the AVs on narrow paths of the road, rather than waiting for the AV to pass. In addition, the types of interaction engaged in by the different road user groups varied considerably, with females and older users more likely to show cautionary behaviour around the AVs than males, or younger road users. Overall, the results highlight the importance of implementing the correct infrastructure to support the safe introduction of AVs, while also ensuring that the behaviour of the AV matches other road users’ expectations as closely as possible in order to avoid traffic conflicts.