Fatigue and mental underload further pronounced in L3 conditionally automated driving: Results from an EEG experiment on a test track
DOI: 10.48550/arxiv.2405.18114
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 transition from manual driving to conditionally automated driving (SAE Level 3, or L3) affects drivers’ mental workload and sleepiness, addressing concerns that passive supervision leads to mental underload and fatigue. While previous research has largely relied on driving simulators, this experiment utilized a realistic test track setting to compare cognitive states across three conditions: manual driving, monitoring an SAE Level 2 (L2) system, and monitoring an L3 system. The research aims to provide empirical data to inform the design of intelligent user interfaces that can mitigate the risks associated with driver inattention and drowsiness in highly automated vehicles. The experiment involved 23 adult participants with no prior automated vehicle experience. Using a Volkswagen Golf prototype on a 2000-meter loop at a test track in Stuttgart, Germany, participants completed three counterbalanced 17-minute drives: manual driving, L2 monitoring (requiring constant supervision), and L3 monitoring (allowing non-driving activities). Mental workload was assessed using the NASA-TLX self-report scale, and sleepiness was measured via the Karolinska Sleepiness Scale (KSS). Additionally, electroencephalography (EEG) data were recorded using 32 channels to analyze oscillatory brain activity. Specifically, the study examined relative spectral power density in the theta band (4–8 Hz) and alpha band (8–12 Hz) at the Fz electrode, as these markers are established indicators of mental workload and fatigue, respectively. The results demonstrated that L3 driving significantly reduced mental workload and increased sleepiness compared to both L2 and manual driving. Self-report data showed lower NASA-TLX scores for L3 rides, with post-hoc tests confirming significant differences against both manual (mean difference = 14.043, p = .002) and L2 (mean difference = 10.957, p = .020) conditions. Similarly, KSS scores indicated higher sleepiness during L3 rides compared to manual driving (mean difference = 0.826, p = .039). EEG analysis corroborated these findings: frontal theta power, associated with mental workload, was significantly lower during L3 rides than during manual driving (mean difference = 0.013, p = .010). Conversely, frontal alpha power, linked to fatigue, was significantly higher during L3 rides compared to manual driving (mean difference = 0.028, p = .016). No significant differences were found between manual and L2 conditions for these metrics. These findings confirm that L3 automation exacerbates mental underload and fatigue compared to lower automation levels, posing safety risks as drivers remain partially responsible for the vehicle. The study concludes that future intelligent user interfaces for L3 vehicles must incorporate dynamic attention management and driving monitoring systems to detect inattention and engage drivers before control transitions are required. The authors emphasize the need for further research in real-world traffic environments and with experienced users to refine these interface designs.
Key finding
Level 3 conditionally automated driving results in significantly lower mental workload and higher sleepiness compared to manual driving and Level 2 automation.
Methodology
on_road
Sample size: 23
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.
| Stage | Outcome | Tool | Model | Prompt | Attempts | Completed |
|---|---|---|---|---|---|---|
| discover | success | — | — | — | 1 | 2026-05-07 |
| 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 | openalex | — | — | 4 | 2026-07-02 |
| 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.
- situational awareness
- drowsiness detection algorithms
- drowsiness
- automation
- automation complacency bias
- mental demand
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: physiological data
- Theoretical Contribution: theory or model, conceptual framework