Interaction between Pedestrians and Autonomous Vehicles in an Urban Environment – Virtual Reality Simulation Study

The introduction of Autonomous Vehicles (AVs) into our current traffic systems represents a significant technological advancement and has the potential to transform urban mobility. However, it also presents the challenge of ensuring a safe environment for all road users. To address this challenge effectively, it is necessary to understand how pedestrians behave in various urban scenarios, which may add an uncertainty. Additionally, the incorporation of AVs requires a level of social acceptance that must coincide with technological advances and legal and regulatory approval [1]. To ensure this, it is vital not only to design the characteristics of autonomous vehicles that are necessary to guarantee a satisfactory level of safety but also those that affect pedestrians' perception of safety. Virtual Reality is crucial in this regard, as it allows researchers to design AVs that are socially aware and study the interaction between AVs and pedestrians. Since AVs are programmed to follow our directives, the aim is to develop better algorithms that mimic human behaviour. However, studying the interaction between AVs and pedestrians in real-life settings is unsafe and difficult. One solution is to conduct Virtual Reality (VR) studies, which allow us to examine the interaction between pedestrians and AVs in various dynamic urban scenarios within a controlled environment. Developing AVs that are socially aware and provide a safe dynamic environment ensures that their driving strategies are perceived as safe by pedestrians. This study focuses on interactions with pedestrians, as they are the most vulnerable group in urban traffic. Walking is the most common and vulnerable mode of transportation, practiced by people of all ages, including children and elderly citizens. Data collected by the European Road Safety Observatory shows that in 2019, pedestrians accounted for 20% of all road fatalities in the EU, and in pedestrian crashes, they comprised 98% of the mortal victims [2]. This underscores the importance of ensuring that autonomous vehicles have driving strategies that are not only safe but also perceived as safe by pedestrians. To achieve this, it is necessary to examine the different forms of communication pedestrians use to interact with other road users. This research will focus on passive forms of communication, how the actions of autonomous vehicles may be interpreted by pedestrians, and how pedestrians will act ac cording to these interpretations. One of the main highlights of the study will be to examine how different stopping maneuvers (quick halt, slow deceleration, early stop, etc.) affect pedestrians' decisions and perception of safety. To this end, a VR study will be conducted where users will be immersed in a scenario encountering an AV implementing different driving and stopping strategies to evaluate their reactions. Thus, the aim of this project is to analyse the features of AVs pertinent to pedestrians' comfort and safety through a VR study. This analysis focuses on deriving the requirements for autonomous vehicles' driving strategies.