Smart IoT based Early Stage Drowsy Driver Detection Management System

This study aimed to develop an innovative alert management system for creating smart cars that detect drowsy driving and prevent it automatically. However, sleepiness is a common physiological occurrence in humans that can occur for various reasons. To prevent the accident's cause, a reliable management system must be designed. In this suggested work, it is proposed that a drowsy driver warning system was created utilising a method where the Video Stream Processing (VSP) used an Eye Aspect Ratio (EAR) and Euclidean distance of the eye to study the eye blink concept. The facial landmark method is also used to distinguish eyes with accuracy. The IoT module delivers a warning message with collision incidence and position information, alarms via voice speaking, and notifies nearby traffic/the owner of the car over the Raspberry Pi tracking system when the driver's fatigue is detected. The suggested model excels in that it can identify tiredness in both daytime and nighttime vision with obstacles at different distances with an accuracy greater than 98%.