Information or advice? A virtual reality study of message content for road crossing in front of autonomous cars

With a view to road safety education and accident prevention, the study of the interaction between autonomous cars and pedestrians cannot be postponed until actual introduction in the public space. In this study, we proactively investigate road-crossing decisions in virtual reality. The objective is to determine the influence of message content displayed by autonomous cars on road crossing decisions and feeling of safety. Focus groups resulted in two different message contents: information about the speed of the car and advice about desired behaviour (pedestrian traffic light). In a virtual reality study, the two message contents were compared to a gear-only control condition and to a human-operated car. The procedure involved 24 crossing decisions in road configurations varying in the likelihood of crossing according to traffic rules. The results showed that road crossing behaviour in virtual reality complied with the likelihood of crossing according to traffic rules (pedestrian rule compliance). Furthermore, as expected, we found an interaction between the type of car and message content in underdetermined road crossing situations. Positive crossings decisions were the least frequent in front of autonomous cars in the pedestrian traffic light (advice) condition. Thus, advice as message content generates cautious pedestrian road crossing behaviour. Finally, the results show a high feeling of safety except when pedestrian traffic lights were absent (present neither as object in the road configuration nor as message content). The results and implications for road safety education are discussed.