With the advent of partially and fully autonomous vehicles, drivers have begun to share more and more control with their cars. However, drivers currently do not have an accurate mental model of autonomous vehicles. This will lead to lower levels of trust and comfort for drivers. Also, drivers’ current mental models towards autonomous cars have not been extensively explored yet. Hence, it is crucial to understand what types of messages and behaviors that a car should perform in order to increase the driver’s confidence and performance. As Norman et al. notes, a
main challenge for automation is the inability to convey the proper information and provide the proper feedback [1]. Will the initiator of an interaction between a driver and a partially or fully autonomous vehicle affect a driver’s performance? When should a car provide information to the driver (push of information / the driver receives the information automatically)? When should a driver request information from the car (pull of information / the driver asks for the information)? What types of information should be exchanged? This research proposal presents a Wizard of Oz study that will be conducted to answer all these questions.