AI-Driven Driver State Monitoring for Road Accident Prevention

Abstract

In recent years, road accidents have become a leading cause of fatalities globally, with driver drowsiness, alcohol consumption, and fatigue playing significant roles in these incidents. To tackle these critical safety issues, this project introduces the design and development of an intelligent Driver Monitoring and Alert system, utilizing a raspberry Pi 4, a USB camera, an MQ2 gas sensor, and a vibration buzzer. The system aims to detect signs of driver fatigue, yawning, and alcohol consumption in real-time and provide immediate alerts to prevent potential accidents. Using computer vision techniques powered by OpenCV and dlib libraries, the system analyses facial landmarks such as eye closure and mouth opening to detect drowsiness and yawning. Additionally, the MQ2 gas sensor monitors alcohol concentration in the driver’s breath, triggering the alert system with high levels of alcohol or prolonged drowsiness are detected. Implemented using Python programming and multithreaded execution for real-time monitoring, the system offers a cost-effective and adaptable solution for integration into both personal and commercial vehicles. The results from extensive testing under various lighting and behavior conditions demonstrate the system’s effectiveness in accurately identifying unsafe driver behaviors with minimal false positives, this project not only enhances road safety but also represent a step toward the development of intelligent, autonomous vehicle monitoring systems. Future improvements could include cloud integration for event logging, GPS module for location tracking during alerts, and AI-based behavior analysis for increased accuracy, contributing to a safer driving environment and reducing accidents caused by human error.