Modelling and Detection of Driver's Fatigue using Ontology

Lambert, Alexandre; Hina, Manolo Dulva; Barth, Celine; Soukane, Assia; Ramdane-Cherif, Amar · 2022 · arXiv

archive: archived pipeline: cataloged verified

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

Summary

This paper addresses the critical safety issue of driver fatigue, a leading cause of fatal road accidents globally. The authors argue that existing fatigue detection models are often unreliable because they rely on a limited subset of symptoms. To improve detection accuracy, the study proposes a comprehensive ontology-based model that integrates measurable data from three distinct sources: vehicular dynamics, driver physical states, and driver physiological signals. The goal is to create an intelligent system capable of continuously inferring fatigue levels and issuing warnings to prevent accidents. The methodology involves constructing a detailed driver fatigue model and representing it using an ontology developed in Protégé using OWL/RDF formats. The model categorizes fatigue indicators into three main classes: vehicle measurements (steering wheel angle, yaw angle, speed, acceleration, and lane position), physical measurements (facial features such as eye closure, blink frequency, mouth condition, and head movement), and physiological measurements (EEG, ECG, EMG, EOG, and skin conductance). The system architecture consists of perception, reasoning, and decision components. Data is acquired via a CARLA driving simulator and sensors, processed into the ontology as instances, and analyzed using Semantic Web Rule Language (SWRL). These rules utilize fuzzy logic to map qualified parameter values (e.g., "High" angular velocity or "Extreme" steering angle) to specific fatigue levels. The findings detail the specific parameters and thresholds used for detection. For vehicular data, the study identifies that steering wheel angles exceeding 6° and high angular velocities (>6°/s) indicate drowsiness, while yaw angle variances correlate with fatigue. Physical indicators include PERCLOS metrics for eye closure and increased blink duration. Physiological markers involve specific EEG frequency bands (increased Alpha/Theta power) and reduced heart rate variability. The paper demonstrates how SWRL rules combine these qualified inputs to infer intermediate and high fatigue states. For instance, a combination of extreme steering wheel angle, high angular velocity, and high correction frequency results in a "High" fatigue classification. The ontology serves as a computer-readable representation that fuses these diverse data streams to determine the driver's state. The significance of this work lies in its holistic approach to fatigue detection, moving beyond single-parameter models to a multi-source data fusion framework. By formalizing fatigue knowledge into an ontology, the system enables robust reasoning and real-time adaptation to driver profiles. This approach enhances the reliability of intelligent vehicle systems, contributing to safer driving environments by providing timely alerts based on a comprehensive analysis of vehicular, physical, and physiological cues. The study establishes a foundational framework for integrating semantic web technologies into automotive safety systems.

Key finding

Multimodal approaches combining physiological signals with behavioral indicators provide the most accurate detection of driver drowsiness, with specific physiological markers showing early warning signs before performance degradation occurs.

Methodology

lab_experiment

Sample size: 20

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. Discovered via discover_arxiv on 2026-05-04 (5 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-07; verification: verified.

Topics

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

• drowsiness detection algorithms
• drowsiness
• truck driver fatigue
• time on task

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
• Theoretical Contribution: computational model