Drowsiness Detection and Alarming System

Girijannavar, Laxmi R; Kallur, Rohini; Kalal, Vijayalaxmi · 2025 · Crossref

DOI: 10.46610/rtsst.2025.v02i01.003

archive: archived pipeline: cataloged verified

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

Summary

This survey paper addresses the critical safety issue of driver drowsiness, which is estimated to cause approximately 30% of all traffic accidents and contribute to over 20% of vehicle crashes. The authors aim to reduce severe injuries, deaths, and economic losses by reviewing existing drowsiness detection techniques. The study categorizes detection methods into five primary approaches: image processing, artificial neural networks (ANN), electroencephalograph (EEG) based systems, vehicular measures, and subjective vocal measures. The objective is to evaluate the efficacy, advantages, and limitations of each method to identify the most efficient system for real-time driver monitoring. The paper conducts a comprehensive literature review of these five techniques. Image processing methods utilize computer vision to analyze facial features, specifically employing eye blinking rates, template matching for open/closed eyes, PERCLOS (percentage of eye closure), and yawning detection based on mouth contour changes. ANN techniques leverage neural networks to process visual cues like eyelid movement, gaze, and head position, with cited research reporting up to 96% accuracy. EEG methods involve wearing electrode helmets to capture brain signals, using power spectrum analysis and FastICA algorithms to determine fatigue levels. Vehicular measures rely on sensors to monitor steering wheel movement (SWM) and standard deviation of lane position (SDLP). Finally, subjective measures utilize the Karolinska Sleepiness Scale (KSS), a nine-point self-rating questionnaire. The findings indicate that no single method provides 100% reliability. Image processing is non-intrusive and easy to implement but suffers from lower accuracy due to sensitivity to lighting conditions and background interference. EEG-based techniques offer high reliability and accuracy but are intrusive, requiring drivers to wear specialized helmets. ANN systems provide accurate results but are computationally intensive and difficult to implement. Vehicular measures, such as SWM and SDLP, are non-intrusive but are unreliable predictors of drowsiness because they are heavily dependent on external factors like road geometry, lighting, and climate, and can be confused with other forms of impaired driving. Subjective measures are useful in simulated environments but are impractical for real-time detection as they require driver introspection, which may artificially reduce drowsiness levels. The authors conclude that a hybrid approach combining multiple detection methods is necessary to overcome the limitations of individual systems. They propose integrating non-intrusive image processing with vehicular measures and physiological indicators, such as heart rate and respiration, potentially using wireless sensors embedded in seat belts. This multi-modal strategy aims to create a robust, non-intrusive, and highly accurate drowsiness detection and alarming system capable of functioning effectively in real-world 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.

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
• dms validation
• distraction detection algorithms
• 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: tool software, validation psychometrics