Edge AI Driver Monitoring System

Description:

Developed a real-time driver monitoring system on the NVIDIA Jetson Orin Nano using Python, OpenCV, and MediaPipe to detect drowsiness and driver inattention from live camera input. The system tracks facial landmarks to identify prolonged eye closure and head-position-based inattention, then publishes structured alert events through an MQTT-first pipeline for cloud relay, database storage, and WebSocket-based downstream consumers. Designed for edge deployment, the project avoids streaming raw video and instead sends lightweight event payloads with metadata, severity levels, timestamps, and detection details, making it suitable for scalable driver safety and ADAS-style monitoring applications.