Abstract
Current state of public reliance has raised concerns about the safety of public transport drivers. Public transport drivers exposed to several risks such as drowsiness, drunk driving and health issues. This study proposes an intelligent system for evaluating risk factors affecting public transport drivers. The proposed system consist of three main components: (1) Drowsiness detection with YOLO based algorithms, which analyzes face signatures for the fatigues; (2) Alcohol intoxication detection, using MQ-3 gas sensor on Raspberry PI based mini computer which will detect alcohol level before vehicle ignition; (3) Heartbeat monitoring using a heartbeat sensor and an MCP3008 analog converter to track drivers heartbeat continuously. The goal is to create real time monitoring system using embedded hardware and machine learning. Current results shows that YOLOv5 model significantly outperformed other models in drowsiness detection problem.