Bringing a Vehicle to a Controlled Stop: Effectiveness of a Dual-Control Scheme for Identifying Driver Drowsiness and Preventing Lane Departures Under Partial Driving Automation Requiring Hands-on-Wheel
DOI: 10.1109/thms.2021.3123171
archive: archived pipeline: cataloged
Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)
Summary
This study addresses the challenge of detecting driver drowsiness and preventing lane departures in hands-on partial driving automation (Level 2), where drivers must keep their hands on the steering wheel. Existing drowsiness detection methods, such as physiological sensors or facial recognition, are often unreliable due to individual variations or occlusions. The authors propose a dual-control scheme that uses vehicle dynamics to simultaneously perform safety control and identify driver state. The system monitors for anticipated lane departures; if detected, it executes a first-stage partial steering control to keep the vehicle within the lane but offset from the center. If the driver fails to correct the vehicle’s position within 10 seconds, the system assumes the driver is not supervising, activates a second-stage control to center the vehicle, and initiates deceleration to bring the car to a controlled stop. The effectiveness of this scheme was evaluated through a fixed-base driving simulator experiment involving 20 participants (14 males, 6 females, aged 20–28). Participants drove a simulated 100 km/h expressway for 60 minutes in a sleep-inducing environment (26°C, early afternoon) to induce passive fatigue. The simulator included a lane centering system (LCS) with torque limits of ±0.5 Nm, which the driver could override. Data were collected at 120 Hz, including vehicle lateral position, steering torque, and eyelid/head pose via cameras. Drowsiness was rated on a 5-point scale by evaluators based on facial expressions and eyelid closure. Statistical analyses included one-way ANOVA to assess the impact of drowsiness levels on driver behavior and Kruskal–Wallis tests to evaluate driver reaction times. Results indicated that while all participants reached the highest drowsiness level (level 5) at least once, the dual-control scheme successfully prevented lane departures in most cases by triggering the first-stage control. However, the system’s ability to accurately and timely identify driver drowsiness remained problematic. The indirect link between drowsiness and controller activation led to issues in the timeliness and accuracy of state identification. Specifically, the system struggled to distinguish between a driver who was merely inattentive and one who was asleep, leading to potential false detections or missed detections. Although the mechanism was effective for safety (preventing departures and executing controlled stops), the driver state identification component required improvement to ensure reliable detection of drowsiness without relying on direct physiological measures. The study concludes that while dual-control schemes are a viable mechanism for enhancing safety in hands-on partial automation by preventing lane departures, they are insufficient for robust driver state identification on their own. The findings highlight the need for integrating more direct driver monitoring methods or improving the logic of state inference to ensure timely intervention. This work contributes to the field of human-machine interaction in automated driving by demonstrating the limitations of using vehicle control actions as proxies for driver state assessment in monotonous driving conditions.
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 | unpaywall | — | — | 2 | 2026-08-09 |
| extract | success | cached | — | — | 4 | 2026-08-23 |
| 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 | success | semantic_scholar | — | — | 1 | 2026-08-09 |
| promote | success | — | — | — | 1 | 2026-08-09 |
| summarize | success | llm | qwen3.8-27b-gittensor | summ-v5 | 3 | 2026-08-23 |
| tag | success | vector_similarity | — | — | 10 | 2026-08-11 |
| verify | success | — | — | — | 1 | 2026-08-09 |
Summary generated by qwen3.8-27b-gittensor on 2026-08-23; verification: pending re-verification.
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).
- Empirical Findings: behavioral performance data, physiological data
- Theoretical Contribution: conceptual framework