Driver Drowsiness Detection with Commercial EEG Headsets
DOI: 10.1109/icrom57054.2022.10025070
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 critical safety issue of driver drowsiness, a leading cause of road accidents, by evaluating the efficacy of commercial electroencephalography (EEG) headsets compared to traditional vehicle-based detection methods. While vehicle-based metrics (e.g., steering angle, lane deviation) are non-intrusive and easy to capture, they often suffer from low accuracy due to external disturbances and delayed symptom manifestation. Conversely, clinical EEG offers high accuracy but is impractical for real-world driving due to its intrusive nature. The authors investigate whether consumer-grade dry-electrode EEG headsets can provide a feasible, high-performance alternative for detecting drowsiness. The experimental design involved 50 volunteers driving a fixed-base simulator under conditions designed to induce drowsiness, including sleep deprivation (averaging 4.5 hours of sleep) and a heavy meal prior to testing. Data were collected from 48 participants after two were excluded due to missing simulator logs. EEG signals were recorded using Muse 2 and Muse S headsets, which utilize four dry electrodes (AF7, AF8, TP9, TP10) and sample at 256 Hz. Drowsiness levels were labeled using the Observer Rating of Drowsiness (ORD) method, where three observers rated video footage of the drivers every 30 seconds on a scale of 1 to 5. Ratings of 1–2 were classified as alert, and 3–5 as drowsy. Vehicle-based features, including steer angle, steer speed, lane deviation, and torque sensor data, were extracted from the simulator logs at 30 Hz. The analysis focused on the statistical separation between alert and drowsy states using p-values derived from Wilcoxon rank-sum tests. For EEG data, absolute and relative power spectral density features were extracted from delta, theta, alpha, beta, and gamma frequency bands. After denoising, 95% of the absolute PSD features showed statistically significant separation (p < 0.05), with beta, theta, and gamma bands proving most effective. In contrast, vehicle-based analysis revealed that while steer angle, lane deviation, and torque sensor data significantly differentiated the states, steer speed did not (p = 0.3590). Crucially, the EEG-based features exhibited lower p-values than the vehicle-based features, indicating a more robust and meaningful separation between alert and drowsy conditions. The study concludes that commercial EEG headsets are feasible and superior to vehicle-based methods for detecting driver drowsiness. The EEG features, particularly absolute power spectral density in specific frequency bands, provide higher sensitivity and accuracy in distinguishing drowsiness states. This finding suggests that consumer-grade EEG devices can overcome the practical limitations of clinical equipment while outperforming the less reliable vehicle-based metrics, offering a promising direction for real-time, non-intrusive driver monitoring systems.
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 | — | — | 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 | success | semantic_scholar | — | — | 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.
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
- Methodological Resource: tool software, validation psychometrics