Impact of Road Event Recognition Reliability in Autonomous Vehicles on Driver Trust and Takeover Performance

Chen, Shan-chih; Liu, Yung-ching · 2025 · Crossref

DOI: 10.54941/ahfe1006732

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 the reliability of road event recognition in autonomous vehicles (AVs) influences driver trust and takeover performance, addressing a critical gap in human-machine interaction research. While AVs promise significant safety improvements by reducing human error, driver trust remains a pivotal factor for safe handovers, particularly as fully autonomous systems lack universal safety certification. The research specifically examines how varying levels of system reliability and different types of recognition errors—false alarms versus misses—affect driver behavior and psychological reliance on the system. The experimental design utilized a driving simulator with 60 participants aged 20–30 who had no prior AV experience. Participants were divided into two groups: one performed a non-driving-related task (NDRT) involving visual search, while the other passively observed. The study employed a mixed factorial design manipulating three reliability levels (93%, 80%, 60%) and two error types (false alarm vs. miss). Each participant completed six standardized 18-minute driving scenarios, encountering 15 recognition events each. Trust was measured via a 100-point scale questionnaire after each event, while driving behavior and takeover reaction times were recorded objectively. Results indicated that miss errors significantly reduced driver trust, increased cognitive load, and delayed subsequent takeover reaction times by approximately 0.41 seconds. In contrast, false alarms had a negligible, non-significant effect, slightly decreasing reaction times by 0.14 seconds. Miss errors also led to unstable driving behaviors, characterized by increased lateral acceleration variability and sudden lane changes. A significant interaction was found between system reliability and secondary task execution; performing an NDRT further decreased trust at lower reliability levels (80% and 60%). Notably, when reliability was 93% and drivers were engaged in an NDRT, the difference in takeover time between hit and miss events widened to 1.57 seconds. Trust did not fully recover to high-reliability levels even after accurate warnings, suggesting persistent skepticism following errors. The findings underscore that omission errors (misses) are more detrimental to trust and safety than commission errors (false alarms). The study concludes that high system reliability is essential for maintaining driver trust, but error management strategies must also account for the compounding effects of cognitive load and secondary tasks. These results imply that AV design should prioritize minimizing miss errors and implementing transparent feedback mechanisms to sustain appropriate trust levels and ensure safe driver takeover in high-risk or distracting environments.

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-prismaquant on 2026-06-10; verification: verified.

Topics

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

• trust calibration
• automation surprise
• takeover transitions
• automation
• trust in automation foundations
• automation complacency bias

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: behavioral performance data, self report data
• Methodological Resource: validation psychometrics