University of Twente Student Theses


Investigating trust calibration during highly automated driving

Steinke, J. (2020) Investigating trust calibration during highly automated driving.

[img] PDF
Abstract:The development of self-driving cars is advancing every day. However, these cars are often perceived as a safety hazard. Possible users often do not know what an automated vehicle is capable of and do not trust the technology. The present study tried to identify when the trust by the user of an automated vehicle is calibrated. In addition, possible improvements for situations with uncalibrated trust were identified. The participants in this study experienced a simulated automated vehicle in various road situations and reported their trust in the vehicle. The trust reported by the participants was then compared with the reliability rating given by engineers of an automated vehicle to identify possible mismatches. A mismatch would cause either undertrust, meaning the user entrusts the vehicle with less than it is capable of, or overtrust, thus the user trusting the vehicle to perform beyond its capabilities. It was found that users often undertrust the vehicle in a situation with poor visibility and overtrust in situations with good visibility. The level of trust was also related to how easy participants perceived the situation which often did not align with the engineers’ perspective. It is advised to use graphical representations of how the vehicle perceives its surroundings in order to calibrate trust. Important information and decisions made by the vehicle should also be presented to the user. Lastly, the simulator used in this study proved to have a comparable effect on trust as a real-life experience.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:55 traffic technology, transport technology, 77 psychology
Programme:Psychology MSc (66604)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page