Developing Driver Safety Monitoring System Using Deep Learning and IOT

The project proposes a Driver Monitoring and Alert System which aims to improve the safety of the driver by enabling real time detection of drowsiness, distraction and unsafe behaviors of the driver. By making use of a Raspberry Pi camera along with YOLOv8, Dlib and MediaPipe the system can accurately detect driver drowsiness, eye closure, yawning and other unsafe behaviors like usage of phone, smoking and consumption of drink; along with detection of driver emotions. If an unsafe act is detected by the system an alarm is generated using the LCD display, buzzer and voice alert, and with the help of the DC motor the speed of the vehicle is simulated to be slowed down or halted entirely. Moreover, the system also sends real-time alerts with images of the detected action pushed to the driver's mobile via Telegram, so they are remotely monitored. This is a low-cost, AI-IoT integrated solution that exploits the benefits of computer vision, real-time alerts and control system to enhance driver safety and alertness. Keywords: Driver Monitoring System (DMS), road safety, driver fatigue detection, drowsiness detection, distraction detection, YOLOv8, Dlib, MediaPipe, Raspberry Pi, computer vision, real-time monitoring, eye closure detection, yawning detection, mobile usage detection, smoking detection, alcohol detection, emotion recognition, AI-IoT integration, alert system, Telegram notifications, automated vehicle control, DC motor simulation, embedded systems, low-cost safety solution.