An Automated System for Driver Drowsiness Monitoring Using Machine Learning

Abstract

Now a days the road accidents are increasing, the primary cause for these accidents is the drowsy driving which leads to death. This is the reason that the detection of driver drowsiness and its indication is foremost in real world. Most of the methods used are vehicle based or Physiological based. Some methods are intrusive and distract the driver, some require expensive sensors manually. But in today’s era real time driver’s drowsiness detection system is very much essential. Hence, the proposed system is developed. In this work a webcam records the video where the driver’s face is detected. Facial landmarks are pointed on the recognized face and the Mouth Opening Ratio(MOR), Eye Aspect Ratio (EAR), and head bending values are calculated, subjective to these values drowsiness is detected based on threshold, and an alarm is given. As the drowsiness is the stage where the driver is unmindful of persons walking on the road, so the pedestrian is detected to avoid any calamity and potholes are identified to avoid the sudden changes in the driving speed which is caused by drowsiness. From the observation, it is found that proposed system works well with 98% of accuracy. If the user is slightly bend for head and mouth then accuracy is less.