🚗 Driver-Safety-Guard: AI-Powered Driver Monitoring System 🛡️

🔍 Overview

Driver-Safety-Guard is an AI-powered driver monitoring system that detects distracted driving behaviors using a YOLOv8 model and integrates with an ESP32 microcontroller to provide real-time alerts via LEDs and a buzzer. This project enhances road safety by identifying signs of drowsiness, mobile phone usage, smoking, and other distractions.

---

🎯 Key Features

✅ Real-Time Detection - Detects driver distractions using a camera and YOLOv8.
✅ ESP32 Integration - Controls LEDs and a buzzer via HTTP requests.
✅ Multiple Alert Mechanisms - Provides visual and sound alerts to warn drivers.
✅ Customizable Model - Train the YOLOv8 model on new datasets for better accuracy.
✅ Cloud & Edge Deployment - Works on local machines and can be adapted for cloud processing.

---

🛠️ Hardware & Software Requirements

🔌 Hardware

• ESP32 Development Board
• Webcam / Camera Module
• LEDs (Green, Red, Orange, Blue)
• Buzzer
• Power Supply (ESP32-compatible)

🖥 Software

• Python 3.11+
• OpenCV (for image processing)
• Ultralytics YOLOv8 (for object detection)
• ESP32 Web Server (for communication)
• Google Colab (for training the model)

---

📂 Project Structure

📂 Driver-Safety-Guard/
 ├── 📁 Model_Training/
 │    ├── Main_Model_Training.ipynb  # Colab Notebook for YOLOv8 training
 │    ├── data.yaml  # Dataset configuration
 │    ├── train/  # Trained model weights
 ├── 📁 ESP32/
 │    ├── Main_Arduino_Code.ino  # ESP32 firmware for LED & Buzzer alerts
 ├── 📁 Model_Run/
 │    ├── Model_Run.py  # Runs the detection and communicates with ESP32
 ├── 📁 Media/
 │    ├── images/  # Sample detection images
 │    ├── presentation.pdf  # Full project presentation
 ├── README.md  # Project documentation

---

🏋️‍♂️ Model Training (YOLOv8)

1️⃣ Mount Google Drive & Extract Dataset

from google.colab import drive
drive.mount('/content/drive')

import zipfile, os
zip_file_path = "/content/drive/MyDrive/dataset.zip"
extracted_dir_path = "/content/extracted"
os.makedirs(extracted_dir_path, exist_ok=True)
with zipfile.ZipFile(zip_file_path, 'r') as zip_ref:
    zip_ref.extractall(extracted_dir_path)

2️⃣ Train the Model

!yolo task=detect mode=train model=yolov8m.pt data=data.yaml epochs=100 imgsz=640 plots=True

3️⃣ Save the Trained Model

import shutil
source_path = "/content/extracted/runs/detect/train"
destination_path = "/content/drive/MyDrive/trained_model"
shutil.copytree(source_path, destination_path, dirs_exist_ok=True)
print("Training results saved.")

---

🔌 ESP32 Setup & Communication

The ESP32 acts as a web server, receiving commands via HTTP requests from the AI model.

1️⃣ ESP32 Web Server Setup

• Connect ESP32 to WiFi.
• Define endpoints (/distracted, /drowsy, /mobileuse, /smoking, /off).
• Control LEDs & Buzzer based on detected behavior.

2️⃣ ESP32 HTTP Endpoint Example

server.on("/distracted", HTTP_GET, [](){
    digitalWrite(greenLED, HIGH);
    tone(buzzer, 1000, 1000);
    server.send(200, "text/plain", "Distracted: Green LED On with Buzzer Beep");
});

---

🚀 Running the Model & ESP32 Integration

1️⃣ Run the Detection Script (Model_Run.py)

python Model_Run.py

2️⃣ ESP32 will Respond to Detected States

• Drowsy → Red LED + Buzzer
• Mobile Use → Orange LED + Buzzer
• Smoking → Blue LED + Buzzer
• Distracted → Green LED + Buzzer

---

📸 Sample Output

Scenario	Detection Preview
Drowsy Driving
Head Drop
Mobile Usage
Smoking

---

📜 License

This project is open-source and licensed under MIT License.

📌 Project Presentation: View Full PDF

🚀 Stay Focused, Stay Safe! 🚗💡

⭐ Star this repository if you like our project! ⭐