Train driver drowsiness detection using deep learning approach

fakhri, parsa; Fathi, Ramin; mirabadi, Ahmad

doi:10.22068/ijrare.344

Volume 11, Issue 1 (6-2024) IJRARE 2024, 11(1): 28-36 | Back to browse issues page

‎ 10.22068/ijrare.344

Mendeley

Zotero

RefWorks

fakhri P, Fathi R, mirabadi A. Train driver drowsiness detection using deep learning approach. IJRARE 2024; 11 (1) :28-36
URL: http://ijrare.iust.ac.ir/article-1-344-en.html

Train driver drowsiness detection using deep learning approach

Parsa Fakhri

, Ramin Fathi

, Ahmad Mirabadi

School of Railway Engineering, Iran University of Science & Technology

Abstract: (56 Views)

Safety is one of the most important priorities of the rail transport industry. Train driver sleepiness is a major threat to railway safety as it can lead to accidents and irreparable losses. In recent years, artificial intelligence and machine learning have emerged as promising approaches for developing driver drowsiness detection strategies. In this article, both deep learning (DL) and convolutional neural network (CNN) approaches are used, and methods to detect train driver drowsiness using YOLOv7 and YOLOv8 models are presented. YOLOv7 and YOLOv8 are the most recent object detection models that are effective for various tasks, including driver drowsiness detection. Our studies show that YOLOv8 outperforms YOLOv7 in terms of accuracy, speed of processing and learning, and required memory in detecting train driver drowsiness. Real-time detection features using YOLOv8 models are also demonstrated. These features can be used to detect drowsiness in real time, which can help prevent accidents.Safety is one of the most important priorities of the rail transport industry. Train driver sleepiness is a major threat to railway safety as it can lead to accidents and irreparable losses. In recent years, artificial intelligence and machine learning have emerged as promising approaches for developing driver drowsiness detection strategies. In this article, both deep learning (DL) and convolutional neural network (CNN) approaches are used, and methods to detect train driver drowsiness using YOLOv7 and YOLOv8 models are presented. YOLOv7 and YOLOv8 are the most recent object detection models that are effective for various tasks, including driver drowsiness detection. Our studies show that YOLOv8 outperforms YOLOv7 in terms of accuracy, speed of processing and learning, and required memory in detecting train driver drowsiness. Real-time detection features using YOLOv8 models are also demonstrated. These features can be used to detect drowsiness in real time, which can help prevent accidents.

Keywords: YOLO, Deep learning, CNN, Train driver, drowsiness detection

Full-Text [PDF 424 kb] (31 Downloads)

Type of Study: Research | Subject: Electrical railway