AI-Based Multimodal Framework for Real-Time Driver Drowsiness Detection and Automated Safety Response

Fuente: Zenodo
Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Albakri, Lilia
Format: Recurso digital
Veröffentlicht: Zenodo 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866901524391657472
author Albakri, Lilia
author_facet Albakri, Lilia
contents <p>Road Guardian is an AI-based multimodal framework for real-time driver drowsiness detection and automated safety response. The system combines facial analysis — Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and head pose estimation via MediaPipe Face Mesh — with Steering Wheel Angle (SWA) monitoring to classify driver state as Normal, Warning, or Critical using a 10-minute temporal sliding window. Upon confirming a Critical unresponsive state, it initiates automated speed reduction, safe pull-over, and emergency contact notification with live GPS coordinates. Built on a lightweight MobileNet CNN, the system runs at 15–30 FPS entirely on-device, projecting 99% visual detection accuracy and a 96% temporal F1-score.</p> <p><em>This work was conducted at Arab International University (AIU), Syria. The official website of the university is: </em><a href="https://www.aiu.edu.sy"><em><span>https://www.aiu.edu.sy</span></em></a></p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_20029486
institution Zenodo
language
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle AI-Based Multimodal Framework for Real-Time Driver Drowsiness Detection and Automated Safety Response
Albakri, Lilia
Driver Drowsiness Detection
Eye Aspect Ratio
Facial Landmark Detection
Convolutional Neural Network
Real-Time Detection
Computer Vision
Road Safety
Driver Monitoring System
MobileNet
PERCLOS
Yawning Detection
Steering Wheel Angle (SWA)
Multimodal Sensor Fusion
Automated Safety Response
Road Safety
<p>Road Guardian is an AI-based multimodal framework for real-time driver drowsiness detection and automated safety response. The system combines facial analysis — Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and head pose estimation via MediaPipe Face Mesh — with Steering Wheel Angle (SWA) monitoring to classify driver state as Normal, Warning, or Critical using a 10-minute temporal sliding window. Upon confirming a Critical unresponsive state, it initiates automated speed reduction, safe pull-over, and emergency contact notification with live GPS coordinates. Built on a lightweight MobileNet CNN, the system runs at 15–30 FPS entirely on-device, projecting 99% visual detection accuracy and a 96% temporal F1-score.</p> <p><em>This work was conducted at Arab International University (AIU), Syria. The official website of the university is: </em><a href="https://www.aiu.edu.sy"><em><span>https://www.aiu.edu.sy</span></em></a></p>
title AI-Based Multimodal Framework for Real-Time Driver Drowsiness Detection and Automated Safety Response
topic Driver Drowsiness Detection
Eye Aspect Ratio
Facial Landmark Detection
Convolutional Neural Network
Real-Time Detection
Computer Vision
Road Safety
Driver Monitoring System
MobileNet
PERCLOS
Yawning Detection
Steering Wheel Angle (SWA)
Multimodal Sensor Fusion
Automated Safety Response
Road Safety
url https://doi.org/10.5281/zenodo.20029486