Highly automated vehicles are increasingly reliable, yet their acceptance also depends on the passenger's psychological comfort. This thesis examines whether a mobile phone used as an external user interface (communicating the vehicle's intentions to the passenger) can increase psychological comfort during highly automated driving under divided attention. To this end, a native iOS application, CarActivity, was developed in Swift and SwiftUI; through the ActivityKit framework it displays the vehicle's current manoeuvre as a Live Activity on the Dynamic Island. The application follows the MVVM architectural pattern and replays a script of 76 events derived from a recording of an actual simulator drive, giving every participant an identical sequence of notifications. The application was evaluated in a within-subjects simulator study (N = 10; eye-tracking N = 8) with interface-present and interface-absent conditions, in which acceptance (van der Laan scale), workload (NASA-TLX) and gaze allocation were measured, complemented by qualitative interviews. Participants responded to the interface positively, but rated it as more useful than pleasant. A visual analysis of the gaze videos and post-drive interviews revealed a clear behavioural difference: with the interface, participants checked the road anticipatorily and in context and reported a greater sense of control, whereas without it they glanced around without context, became absorbed in the secondary task, and often noticed events only after they had passed. The value of the interface thus lies mainly in the passenger's sense of control and situational awareness, and its effect is strongly individual in degree. The thesis contributes a working external-interface prototype and a mixed-methods evaluation of it.
|
