Effects of Take-Over Requests and Cultural Background on Automation Trust in Highly Automated Driving

Hergeth, Sebastian; Lorenz, Lutz; Krems, Josef F; Toenert, Lars · 2015 · Crossref

DOI: 10.17077/drivingassessment.1591

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 how Take-Over Requests (TORs) and cultural background influence automation trust in Highly Automated Driving Systems (HADS). While HADS can enhance safety and comfort, their effectiveness relies on appropriate driver trust. TORs are necessary due to imperfect system reliability, but their impact on trust is ambiguous; they may be perceived as system failures that decrease trust or as useful warnings that maintain it. Additionally, cultural differences, particularly between Western and Asian drivers, have been cited as potential factors in automation trust, though this had not been previously examined in the context of HADS. The researchers hypothesized that TORs would affect trust and that Chinese drivers would exhibit higher trust in HADS than German drivers. The study employed a driving simulator experiment with 80 participants (40 German, 40 Chinese), all employees of the BMW Group. The design was a two-factor mixed between-within subjects study, with cultural background as the between-subjects factor and time of measurement as the within-subjects factor. Trust was measured using the Automation Trust Scale (pre- and post-session) and single-item trust ratings collected eight times during the session. Behavioral measures included take-over times and resumption lags (time to reactivate the system after a TOR). Participants engaged in non-driving related tasks while the HADS controlled the vehicle. Two TORs were triggered by simulated accidents, requiring drivers to manually take control. Results indicated that TORs temporarily lowered single-item automation trust ratings, suggesting they were initially perceived as automation failures. However, overall trust increased significantly from the beginning to the end of the session for both groups, implying that experiencing the system’s performance and understanding its limitations fostered long-term trust. There were no significant differences in overall automation trust between German and Chinese drivers. However, Chinese participants reported significantly higher automation mistrust than German participants both before and after the experiment. Notably, self-report measures of trust did not correlate with behavioral measures; trust ratings did not predict take-over times or resumption lags. The findings support distinguishing automation trust and mistrust as separate constructs and highlight the dual nature of TORs: they cause short-term trust dips but contribute to long-term trust formation by clarifying system boundaries. The lack of correlation between self-reported trust and behavior suggests that current behavioral metrics may not reliably capture trust in HADS contexts, possibly due to constrained reaction windows. The study concludes that while cultural background influences mistrust levels, it does not significantly alter the general trajectory of trust formation in HADS. Future research should explore different TOR types and long-term trust development.

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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• trust calibration
• trust in automation foundations
• automation
• automation surprise
• takeover transitions
• acceptance adoption

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