Is conditionally automated driving a bad idea? Observations from an on-road study in automated vehicles with multiple levels of driving automation
DOI: 10.1016/j.apergo.2025.104617
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 whether conditionally automated driving (SAE Level 3) presents significant usability challenges when integrated into vehicles that also offer lower levels of automation (Manual and SAE Level 2). The research is motivated by the increasing market availability of Level 3 systems and the resulting complexity in human-machine interaction, particularly regarding "mode awareness"—the driver’s understanding of the current automation state and their corresponding responsibilities. The authors aim to identify specific interaction errors and confusion drivers experience when managing multiple automation levels in a real-world setting. The researchers conducted an on-road study using a Wizard-of-Oz vehicle, a modified Volvo XC90, to simulate Manual, Level 2, and Level 3 driving automation. Sixteen participants drove on public highways in Gothenburg, Sweden. In this setup, participants manually controlled the vehicle or operated the Level 2 assistive system, while a hidden "driving wizard" simulated the unsupervised Level 3 automation. Data collection involved video recordings of driver behavior and steering wheel interactions, think-aloud protocols during driving, and semi-structured post-driving interviews. The study utilized two different user interface designs to present mode states, though the primary focus was on identifying common interaction challenges across both interfaces rather than comparing their efficacy. The results revealed significant errors in mode transitions and mode awareness. Drivers frequently experienced "transition challenges," including the unintended deactivation of Level 2 automation and attempts to activate Level 3 outside its operational design domain. Furthermore, participants exhibited "mode confusion," characterized by uncertainty about their responsibilities after disengaging Level 3 or confusion regarding the active driving mode. For instance, drivers sometimes failed to recognize that they were responsible for primary driving tasks when the system was not active, or they misinterpreted the system's state despite UI indications. These errors persisted regardless of the specific user interface used, suggesting that the fundamental issue lies in the cognitive difficulty of distinguishing between and transitioning across multiple automation modes within a single vehicle. The study concludes that the current design of automated vehicles with multiple automation levels poses a fundamental limitation for human factors. The ambiguity in mode states and the lack of clear mental models among drivers lead to inappropriate responses and safety risks. The findings imply that simply improving user interfaces is insufficient; instead, the design of automation systems must better align with natural human cognitive processes to mitigate mode confusion. The authors recommend that future vehicle designs prioritize clearer differentiation between automation levels and more intuitive transition logic to ensure drivers maintain accurate mode awareness and appropriate behavioral responses.
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 scout_discovery on 2026-05-08 (2 acquisition events logged).
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | partial | scout | — | — | 2 | 2026-05-08 |
| archive | success | canonical_url | — | — | 12 | 2026-06-09 |
| extract | success | cached | — | — | 2 | 2026-06-09 |
| 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 | 2026-05-08 |
| promote | success | — | — | — | 1 | 2026-06-04 |
| summarize | success | llm | qwen3.6-27b-prismaquant | summ-v5 | 1 | 2026-06-09 |
| tag | success | vector_similarity | — | — | 15 | 2026-06-11 |
| verify | success | — | — | — | 1 | 2026-06-09 |
Summary generated by qwen3.6-27b-prismaquant on 2026-06-09; verification: verified.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- mode awareness
- automation
- automation surprise
- situational awareness
- automation complacency bias
- driverless ads
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).
- Theoretical Contribution: conceptual framework