Using pupillometry and gaze-based metrics for understanding drivers’ mental workload during automated driving
DOI: 10.1016/j.trf.2023.02.015
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 drivers’ mental workload during automated driving, specifically comparing SAE Level 2 (L2) and Level 3 (L3) automation scenarios. The research addresses the need for objective, non-invasive indicators of driver state to ensure safety during transitions from automated to manual control. While previous studies often relied on subjective ratings or heart-rate measures, this work focuses on pupillometry and gaze-based metrics, specifically pupil diameter, to capture moment-to-moment fluctuations in cognitive load. The study examines how workload varies during automated car-following (ACF), manual car-following (MCF), and takeover requests, manipulating variables such as time headway (THW) and the presence of a lead vehicle. The experiment utilized a driving simulator with 32 participants divided into L2 and L3 groups. L2 drivers monitored the road during automation, while L3 drivers performed a non-driving related task (NDRT). Each participant completed two experimental drives involving ACF segments with either short (0.5 s) or long (1.5 s) THWs, followed by takeover requests with or without a lead vehicle. Eye-tracking data, including mean and standard deviation of pupil diameter, were collected alongside self-reported workload ratings. Data analysis involved repeated measures ANOVA to compare workload levels across drive modes and takeover phases, segmenting data into pre-takeover, takeover, and post-takeover windows. Results indicated that while mean pupil diameter did not significantly differ between ACF and MCF in the L2 group, the standard deviation of pupil diameter was significantly higher during ACF, suggesting greater fluctuation in mental workload during monitoring compared to manual driving. During takeover requests, mean pupil diameter increased steeply, indicating a sharp rise in workload. This increase was exacerbated by the presence of a lead vehicle, particularly in short THW conditions, for both L2 and L3 groups. Pupil diameter metrics closely mirrored trends in self-reported workload ratings, validating their sensitivity to subtle variations in mental load. The study found that prior engagement in an NDRT (L3) did not significantly alter the workload spike during takeovers compared to monitoring (L2), though the presence of a lead vehicle consistently heightened workload. The findings demonstrate that pupil diameter, particularly its standard deviation, is a sensitive indicator of phasic fluctuations in mental workload during automated driving. The study concludes that eye-tracking metrics can effectively capture real-time changes in driver cognitive load, especially during critical transitions like takeovers. These results support the feasibility of integrating pupillometry into future driver state monitoring systems to assess readiness and safety during automated driving. However, the authors note that further research is needed to validate these metrics in real-world settings and in combination with other physiological sensors.
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 | Crossref | — | — | 1 | 2026-08-09 |
| archive | success | openalex | — | — | 5 | 2026-08-09 |
| extract | success | cached | — | — | 3 | 2026-08-10 |
| clean | success | clean | — | — | 1 | 2026-08-09 |
| chunk | success | chunk | — | — | 1 | 2026-08-09 |
| embed | success | embed | Qwen/Qwen3-Embedding-8B | — | 1 | 2026-08-09 |
| enrich | failed | — | — | — | 1 | 2026-08-09 |
| promote | success | — | — | — | 1 | 2026-08-09 |
| summarize | success | llm | qwen3.6-27b-nvidia | summ-v5 | 2 | 2026-08-10 |
| tag | success | vector_similarity | — | — | 10 | 2026-08-11 |
| verify | partial | — | — | — | 2 | 2026-08-10 |
Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified_with_issues.
Topics
Ranked by relevance to this paper. Hover a topic for its definition.
- workload measurement
- mental demand
- temporal
- hands on hands off engagement
- situational awareness
- automation
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, behavioral performance data
- Theoretical Contribution: theory or model