Design of ADAS Fatigue Control System using Pynq z1 and Jetson Xavier NX

Abstract

In recent years, driver fatigue and drowsiness have been presented as the main factors of road accidents with a high rate of passenger deaths, injuries and property losses. In this paper, we developed a monitoring system for vehicle drivers that detects and warns of the presence of fatigue and drowsiness through computer vision or machine vision. Our developed system consists of four steps: training of the classifier to perform the detection module, image acquisition and processing, detection and alarm activation. Firstly, training of classifier was performed using Haar technique to locate face features. Secondly, a Jetson Xavier NX or Pynq Z1 development board was used to perform real-time video using a 4K camera, thirdly, we use the training classifier to detect the face image and facial landmark algorithm proposed by Kazemi & Sullivan to detect face features. Finally, our system analyses the state characteristics of the eyes; once the processing is completed, the detection of drowsiness and fatigue is carried out.