A Hybrid Approach to Detect Driver Drowsiness Utilizing Physiological Signals to Improve System Performance and Wearability

Awais, Muhammad; Badruddin, Nasreen; Drieberg, Micheal · 2017 · Crossref

DOI: 10.3390/s17091991

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 major contributor to fatal road accidents, by proposing a hybrid detection method that integrates electrocardiography (ECG) and electroencephalography (EEG) signals. The research is motivated by the limitations of existing detection systems, which often rely on single-modal physiological measures or non-intrusive behavioral metrics that lack reliability in real-world conditions. Specifically, while EEG is considered highly reliable for drowsiness detection, its practical application is hindered by poor wearability due to the need for numerous electrodes. The authors aim to improve both detection performance and system wearability by combining EEG and ECG features and investigating the feasibility of reducing the number of required sensors. The experimental design involved 22 healthy university students participating in a simulator-based driving study. Drowsiness was induced using a monotonous driving environment lasting 80 minutes. Physiological data were collected using a dry-electrode Enobio-20 channel device, capturing both EEG and ECG signals. Ground truth for drowsiness states was established through video analysis, where raters identified drowsy events based on facial features such as eye blink duration and yawning. Only data from the 11 subjects who exhibited drowsiness were used for analysis, with 5-minute windows preceding and following drowsy events labeled as alert and drowsy states, respectively. Feature extraction included time-domain statistical descriptors and complexity measures (sample entropy) from EEG, as well as heart rate and heart rate variability (HRV) metrics from ECG. Frequency domain analysis of EEG covered delta, theta, alpha, beta, and gamma bands. Paired t-tests were employed to select statistically significant features (p < 0.05), which were then classified using a support vector machine (SVM). The results demonstrated that combining EEG and ECG features significantly improved the system’s ability to discriminate between alert and drowsy states compared to using either modality alone. A key finding was the successful implementation of a channel reduction strategy. The analysis revealed that an acceptable accuracy level of 80% could be achieved by utilizing only two electrodes: one for EEG and one for ECG. This finding suggests that high-performance drowsiness detection does not require the extensive electrode arrays typical of traditional EEG systems. The significance of this work lies in its contribution to the development of practical, wearable driver monitoring systems. By proving that a hybrid approach with minimal sensors can maintain high accuracy, the study offers a viable solution that balances detection reliability with user comfort. This advancement addresses the wearability barrier that has previously limited the real-world deployment of physiological-based drowsiness detection systems, potentially leading to more effective and widely adopted safety technologies in transportation.

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.

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.

• drowsiness detection algorithms
• drowsiness
• microsleep
• dms validation
• truck driver fatigue
• workload measurement

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: validation psychometrics, tool software