A Systematic Evaluation of Solutions for the Final 100m Challenge of Highly Automated Vehicles

Abstract

Automated vehicles will change the interaction with the user drastically. While freeing the user of the driving task for most of the journey, the "final 100 meters problem'', directing the vehicle to the final parking spot, could require human intervention. Therefore, we present a classification of interaction concepts for automated vehicles based on modality and interaction mode. In a subsequent Virtual Reality study (N=16), we evaluated sixteen interaction concepts. We found that the medially abstracted interaction mode was consistently rated most usable over all modalities (joystick, speech, gaze, gesture, and tablet). While the steering wheel was still preferred, our findings indicate that other interaction concepts are usable if the steering wheel were unavailable.