A LOW-COST AIOT-BASED DRIVER DROWSINESS DETECTION AND REAL-TIME ALERT SYSTEM USING MEDIAPIPE AND ESP32

This paper proposes a low-cost AIoT-based driver drowsiness detection and real-time alert system using MediaPipe Face Mesh and ESP32 embedded hardware. The proposed system integrates computer vision, embedded systems, and Internet of Things (IoT) technologies to detect fatigue behaviors such as eye closure and yawning. The system adopts Python, OpenCV, and MediaPipe to extract 468 facial landmarks in real time and employs Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR) algorithms to determine fatigue conditions. Furthermore, a three-level warning mechanism consisting of LED indicators, buzzer alarms, and Discord Webhook cloud notifications is implemented using ESP32. Experimental results demonstrate that the proposed system achieves real-time processing performance above 30 FPS with low latency and stable operation under different illumination conditions. The proposed architecture provides a practical and scalable solution for intelligent transportation and vehicle safety applications.