Enhancing the Driver's Comprehension of ADS's System Limitations: An HMI for Providing Request-to-Intervene Trigger Information

Matsuo, Ryuji; Liu, Hailong; Hiraoka, Toshihiro; Wada, Takahiro · 2023 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2306.01328

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 addresses the challenge of driver comprehension regarding the system limitations of Level 3 automated driving systems (ADS). When an ADS exceeds its operational design domain, it issues a request-to-intervene (RtI), requiring the driver to take control. However, complex traffic scenarios often present multiple potential triggers (e.g., fog and curves simultaneously), causing driver hesitation or confusion about the specific cause of the RtI. To mitigate this, the authors propose a voice-based human-machine interface (HMI) that provides specific trigger cues during the RtI and post-event reason explanations to help drivers build an accurate mental model of ADS limitations. The researchers conducted a between-group experiment using a driving simulator with 20 licensed participants, divided into two groups: one receiving the proposed HMI with trigger cues and reason explanations, and a control group receiving conventional RtI alerts without such details. The ADS was programmed with specific limitations: RtI was triggered if the road curvature radius was less than 230 meters or if visibility dropped below 40 meters due to fog. Participants underwent four experimental phases involving various combinations of fog and curves, including scenarios with multiple potential triggers and a final "RtI failure" scenario where the system disengaged without warning. Performance was measured via a post-experiment comprehension test, take-over timing during the failure scenario, and collision rates. The results demonstrated that the proposed HMI significantly enhanced driver understanding. In the comprehension test, the group using the trigger cue and reason HMI achieved a median score of 13 out of 13, significantly higher than the control group’s median of 8. During the RtI failure scenario, participants in the experimental group took over control significantly earlier (active take-over) compared to the control group, who often reacted after the ADS had already disengaged (inactive take-over). While the experimental group had fewer collisions (one out of ten) compared to the control group (five out of ten), this difference was not statistically significant. The study concludes that providing explicit trigger cues and reason explanations via voice HMI helps drivers correctly identify ADS limitations and respond more proactively during take-over events, particularly when the system fails to issue a warning. This approach fosters a more accurate mental model of the ADS, reducing hesitation and improving safety during critical transitions from automated to manual driving. The findings suggest that enhancing driver comprehension of system boundaries is crucial for the safe commercialization of Level 3 automation.

Key finding

Providing voice-based trigger cues and reason explanations during request-to-intervene events significantly improves drivers' comprehension of ADS system limitations and leads to faster, more proactive take-over responses in failure scenarios.

Methodology

simulator

Sample size: 20

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.

StageOutcomeToolModelPromptAttemptsCompleted
discover success author_sweep 2 2026-05-28
archive success canonical_url 1 2026-06-04
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).