The Veronica project data includes two parts: taxi passenger choice data and pedestrian crossing data. This part primarily involves pedestrian crossing data. The key motivation for the study is to measure the pedestrain crossing behaviours in the mixed traffic of autonomous vehicles (AVs) and normal vehicles. This study uses virtual reality (VR) to simulate two urban mid-block environments, one in downtown Toronto, Canada and the other in central Newcastle, UK, to investigate the crossing behaviour of 428 participants (9,262 observations). The research questions addressed in this paper, in the context of unmarked mid-block crossing, are: (a) Do various vehicle types, i.e., normal vehicles and AVs, impact pedestrian behaviour? (b) How do other pedestrians influence one’s crossing behaviour? (c) How do traffic characteristics, road type, and environmental characteristics impact pedestrian behaviour? (d) What is the influence of demographics on pedestrian behaviour? and, Are there differences in pedestrian behaviour in different countries?Many governments worldwide are introducing new plans to promote and anticipate the recent rapid development of automation. The economic and societal benefits of autonomous vehicles are foreseen to be enormous (up to Euros 17tn to GDP). But, these benefits could be jeopardised if users fail to adopt the technology. In response to this urgent need, the project aims to take advantage of virtual reality technologies to use them in a scientific context to understand and then model users' acceptance of Fully Autonomous Vehicles (FAVs), particularly Fully Automated Taxis (FATs). In order to achieve the research overall aim, the project set the following specific objectives:- Understand to which extent the acceptance of FATs is affected by how much familiar we can get with this highly technological and innovative product and by what other people around us think about and how they behave with respect to FATs;- Develop a new method for studying acceptance of innovative products, which includes a method to collect the information from customers and a method to analyse this data that can be used then to take policy and industrial decisions that affect every day citizens' life;- Test the benefit of the experiment created to be used across the population to help people to live in and adapt to the forthcoming new technological urban environments. This research will add extensive value to the critical discussions about adoption and diffusion of FAVs or FATs and about the policy incentives that should be given to foster the market. The usability of Virtual Reality environment to social contexts will open opportunities for new applications.

The study includes field virtual reality (VR) experiment using the VR program ‘Unity’. The experimental design was executed with consideration of all research variables listed in the first sheet and practical factors, including the maximum VR immersion duration and the overall experiment duration. In total 12 different sessions were constructed for each participant to go through and each session was repeated twice. The initial combinations were tested with pilot participants to evaluate the performance of the VR experiment and observe participant’s reactions to different environments. All variable levels were incorporated in the final experimental design randomly, having avatar behaviour levels being repeated equally for each participant. A random sampling of the variable levels was performed for each session. Participants were placed on a sidewalk and they were instructed to go to the tactile paving and then find a safe and suitable moment to cross the street. Participants were given a 60-second time limit to complete the task, and if they did not finish within that time frame, the session would be terminated and the next session would be automatically loaded. In Newcastle, 3% sessions resulted in failure to cross, while in Toronto it was 10%. Overall, 171 and 257 individuals participated in the experiment in Toronto and Newcastle, respectively, resulting in 9,262 (≈ (171 + 257) × 12 × 2) observations after removing the unsuccessful crossings. The majority of participants were recruited by a panel provider to retain population profiles and they were compensated.