Real-time and Efficient EEG Based Driver Drowsiness Detection System with VANET for Safe Driving

Mohammedi, Mohamed; Mokrani, Juba; Mouhoubi, Abdenour · 2023 · Crossref

DOI: 10.21203/rs.3.rs-1972742/v1

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the critical safety issue of vehicular accidents caused by driver drowsiness, a significant contributor to fatalities globally. While existing drowsiness detection systems rely on physiological signals, vehicle parameters, or facial behavior, many lack the capability to alert nearby vehicles, limiting their effectiveness in preventing collisions. The authors propose a real-time, efficient drowsiness detection system that integrates Electroencephalography (EEG) signal analysis with Vehicular Ad hoc Networks (VANETs). The primary motivation is to create an automated system that not only detects early signs of drowsiness in the driver but also disseminates this information to surrounding vehicles and roadside infrastructure to enhance overall road safety. The proposed system utilizes a lightweight wireless headset to acquire EEG signals from the C3-O1 channel, identified as optimal for monitoring drowsiness. The methodology involves four phases: signal acquisition, preprocessing, feature extraction, and classification. Raw EEG data is preprocessed using a second-order Butterworth bandpass filter (0.5–30 Hz), Independent Component Analysis (ICA) for artifact removal, and Discrete Wavelet Transform (DWT) with Daubechies 4 wavelets for multi-resolution analysis. Feature extraction computes the Relative Power Ratio (RPR) of alpha and beta waves from 2-second epochs. These features are classified using a linear Support Vector Machine (SVM). If the SVM output indicates drowsiness, the system triggers an on-board alarm and broadcasts an alert via VANETs to nearby vehicles and Roadside Units (RSUs). The system was validated using the MIT-BIH Polysomnographic dataset, which contains EEG recordings from 11 subjects, focusing on transitions between wakefulness and drowsiness (S1 stage). Experiments were conducted in a Matlab/Simulink environment. The study evaluated performance using metrics including accuracy, recall, precision, specificity, and F-score. The results demonstrate that the proposed system achieves high accuracy and robustness in detecting drowsiness states. The integration of VANETs allows for the real-time dissemination of drowsiness alerts, enabling other drivers to take precautionary measures. The significance of this work lies in its dual approach to safety: individual driver alerting and community-wide warning through vehicular communication. By leveraging EEG signals for direct physiological monitoring and VANETs for data dissemination, the system offers a more comprehensive solution than embedded-only models. The authors conclude that the system is feasible as a subjective detection tool for drowsy level analysis, offering improved detection resolution and early warning capabilities compared to existing methods. This approach contributes to the field by bridging the gap between individual physiological monitoring and cooperative intelligent transport 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.

Summary generated by qwen3.6-27b-nvidia on 2026-08-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• drowsiness detection algorithms
• drowsiness
• microsleep

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