A Case Study of Trust on Autonomous Driving
DOI: 10.1109/itsc.2019.8917251
archive: archived pipeline: cataloged verified
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study investigates the dynamic evolution of human trust during interactions with autonomous vehicles, addressing the limitation of existing methods that rely on post-experiment surveys or group-level physiological analysis. The authors argue that real-time, individualized trust assessment is critical for safety, as undertrust leads to automation neglect while overtrust causes delayed intervention. To capture these dynamics, the researchers conducted a human subjects study involving 19 participants in a high-fidelity driving simulator. The experimental design manipulated three independent variables: alarm type (all activated, all missing, early false alarm, late false alarm), weather conditions (sunny vs. rainy), and driving mode (fully autonomous vs. semi-autonomous). Participants rated their trust on a 5-point Likert scale in real-time using steering wheel buttons while navigating 16 scenarios containing hazardous events. Simultaneously, physiological data, including galvanic skin response (GSR), heart rate variability (HRV), and pupil size, were collected as complementary indicators. The researchers employed Analysis of Variance (ANOVA) to assess group-level effects and Signal Temporal Logic (STL) to analyze individual temporal patterns of trust evolution. ANOVA results indicated that alarm type significantly influenced average trust levels, with missing alarms resulting in significantly lower trust compared to activated or false alarms. Weather conditions and driving mode showed weak or non-significant effects on trust. Physiological metrics revealed that pupil size and GSR peaks were sensitive to weather conditions, while HRV varied significantly by gender. STL analysis provided granular insights into trust dynamics: trust was more likely to increase during periods of uneventful driving. Regarding alarms, missing alarms had a greater negative impact on trust than false alarms; specifically, 67.1% of trials with missing alarms resulted in decreased trust within 10 seconds, compared to 51.3% for activated alarms. Furthermore, STL learning techniques identified significant individual variability in reaction times, with some participants adjusting trust levels much faster than others. The study concludes that while ANOVA provides generalized trends, STL is essential for capturing individualized trust patterns and temporal behaviors. The findings highlight that system reliability, particularly the absence of expected alarms, critically undermines user trust. The authors suggest that future autonomous systems should incorporate real-time trust monitoring and personalized adaptation strategies. However, they note limitations regarding the small sample size and the specific demographic of engineering students, recommending further research with more diverse populations to generalize these findings.
Key finding
Missing alarms significantly decreased user trust more than false alarms, while weather conditions and driving modes had minimal impact on trust levels.
Methodology
simulator
Sample size: 19
Provenance
The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | author_sweep | — | — | 2 | 2026-05-28 |
| archive | success | canonical_url | — | — | 1 | 2026-06-06 |
| extract | success | cached | — | — | 3 | 2026-06-10 |
| clean | success | clean | — | — | 1 | 2026-06-04 |
| chunk | success | chunk | — | — | 1 | 2026-06-04 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-06-04 |
| enrich | success | — | — | — | 1 | 2026-05-28 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 2 | 2026-06-10 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 2 | 2026-06-10 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
Information type
What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).
- Empirical Findings: self report data, physiological data, behavioral performance data