Use this URL to cite or link to this record in EThOS:
Title: Enhancing user experience and safety in the context of automated driving through uncertainty communication
Author: Kunze, Alexander
ISNI:       0000 0004 8500 7948
Awarding Body: Loughborough University
Current Institution: Loughborough University
Date of Award: 2019
Availability of Full Text:
Access from EThOS:
Access from Institution:
Operators of highly automated driving systems may exhibit behaviour characteristic of overtrust issues due to an insufficient awareness of automation fallibility. Consequently, situation awareness in critical situations is reduced and safe driving performance following emergency takeovers is impeded. Previous research has indicated that conveying system uncertainties may alleviate these issues. However, existing approaches require drivers to attend the uncertainty information with focal attention, likely resulting in missed changes when engaged in non-driving-related tasks. This research project expands on existing work regarding uncertainty communication in the context of automated driving. Specifically, it aims to investigate the implications of conveying uncertainties under consideration of non-driving-related tasks and, based on the outcomes, develop and evaluate an uncertainty display that enhances both user experience and driving safety. In a first step, the impact of visually conveying uncertainties was investigated under consideration of workload, trust, monitoring behaviour, non-driving-related tasks, takeover performance, and situation awareness. For this, an anthropomorphic visual uncertainty display located in the instrument cluster was developed. While the hypothesised benefits for trust calibration and situation awareness were confirmed, the results indicate that visually conveying uncertainties leads to an increased perceived effort due to a higher frequency of monitoring glances. Building on these findings, peripheral awareness displays were explored as a means for conveying uncertainties without the need for focused attention to reduce monitoring glances. As a prerequisite for developing such a display, a systematic literature review was conducted to identify evaluation methods and criteria, which were then coerced into a comprehensive framework. Grounded in this framework, a peripheral awareness display for uncertainty communication was developed and subsequently compared with the initially proposed visual anthropomorphic uncertainty display in a driving simulator study. Eye tracking and subjective workload data indicate that the peripheral awareness display reduces the monitoring effort relative to the visual display, while driving performance and trust data highlight that the benefits of uncertainty communication are maintained. Further, this research project addresses the implications of increasing the functional detail of uncertainty information. Results of a driving simulator study indicate that particularly workload should be considered when increasing the functional detail of uncertainty information. Expanding upon this approach, an augmented reality display concept was developed and a set of visual variables was explored in a forced choice sorting task to assess their ordinal characteristics. Particularly changes in colour hue and animation-based variables received high preference ratings and were ordered consistently from low to high uncertainty. This research project has contributed a series of novel insights and ideas to the field of human factors in automated driving. It confirmed that conveying uncertainties improves trust calibration and situation awareness, but highlighted that using a visual display lessens the positive effects. Addressing this shortcoming, a peripheral awareness display was designed applying a dedicated evaluation framework. Compared with the previously employed visual display, it decreased monitoring glances and, consequentially, perceived effort. Further, an augmented reality-based uncertainty display concept was developed to minimise the workload increments associated with increases in the functional detail of uncertainty information.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
Keywords: Automated driving ; uncertainty ; trust in automation ; situation awareness ; interaction design ; human-machine interface