Novel Interfaces for Fostering Trust in Autonomous Vehicles

Abstract: This thesis presents the exploration of two interfaces for autonomous vehicles. Autonomous vehicles likely will not be trusted by the general public. By showing intermediate results of the automation, users should be able to feel more trust in the system. Our interfaces used LED lights, the other augmented reality, in order to signal the human of detected objects in the proximity of the vehicle. Prototypes of both interfaces were developed in two different simulator environments, and tested in a series of user experiments. The two simulator environments form an extendable platform for future work. Both interfaces are found to increase trust in the vehicle, the augmented reality interface increased trust the most. Demographic preferences were studied as well, and a small preference for the augmented reality interface by less experienced drivers was found. In further experiments, it was found that the LED interface works better in extreme weather conditions while the augmented reality interface performed better in clear conditions.