🚗 Driver Safety & Alert System (DSAS)

A production-grade driver monitoring solution engineered to prevent accidents caused by drowsiness, fatigue, and distraction. This system leverages computer vision and real-time biometric analysis to continuously assess driver state and trigger multi-modal alerts (Audio, Visual, Email) in critical situations.

Unlike basic prototypes, DSAS features a robust modular architecture, gesture-based SOS system, and a professional analytics dashboard, making it suitable for fleet management integration and advanced safety applications.

---

📸 System Capabilities

1. Intelligent Drowsiness Detection

• PERCLOS Analysis: Monitors percentage of eye closure over time using EAR (Eye Aspect Ratio).
• Micro-Sleep Detection: Identifies brief, involuntary sleep episodes (0.5s - 2s).
• Yawn Frequency: Tracks mouth opening duration (MAR) to detect early signs of fatigue.

2. Attention Monitoring

• Head Pose Estimation: Calculates 3D yaw, pitch, and roll to detect when the driver looks away from the road.
• Distraction Alerts: Triggers immediate warnings for sustained inattention (>2s).

3. 🆘 SOS Emergency Protocol

• Gesture Recognition: Detects a specific "Hollow Namaste" hand gesture held for 5 seconds.
• Automated Dispatch: Instantly sends an SMTP Email Alert to emergency contacts.
• Rich Metadata: Includes GPS coordinates, timestamp, Google Maps link, and driver details.
• Spam Prevention: Built-in 5-minute cooldown to prevent duplicate emergency broadcasts.

4. Real-Time Analytics Dashboard

• Visual Feedback: Live graphs for EAR (Eye Aspect Ratio) and Fatigue Scores.
• Session Metrics: Tracks blink rate, microsleep events, and attention lapses.
• Dynamic UI: Modern, dark-themed interface designed for low-light vehicle environments.

---

🏗️ Technical Architecture

The project follows a Service-Oriented Architecture (SOA) pattern, ensuring separation of concerns and scalability.

Driver_Safety_Dashboard/
├── core/                  # 🧠 Intelligent Processing Units
│   ├── face_monitor.py    # Facial Landmarks & State Machine (EAR/MAR/Pose)
│   ├── hand_gesture.py    # Hand Tracking & SOS Logic (MediaPipe)
│   ├── alert_manager.py   # Priority Queueing & Alert Orchestration
│   └── analytics_engine.py# Statistical Analysis & Scoring Algorithms
│
├── services/              # 🔌 External Integrations
│   ├── email_service.py   # SMTP Client (Gmail/Outlook) with HTML Templates
│   ├── audio_service.py   # Multi-threaded Audio/TTS Engine
│   └── location_service.py# IP-based Geolocation & Fallback Logic
│
├── ui/                    # 🎨 Visualization Layer
│   ├── dashboard.py       # Analytics Rendering Engine
│   ├── display.py         # Main Video Compositor & Overlay Manager
│   └── themes.py          # Centralized Design System (Colors/Fonts)
│
├── config/                # ⚙️ Configuration Management
│   ├── settings.json      # System Thresholds & API Keys
│   └── contacts.json      # Emergency Contact Database
│
└── assets/                # 📦 Static Resources
    ├── models/            # Quantized TFLite Models (Face/Hand)
    └── sounds/            # High-fidelity WAV Alerts

---

🚀 Key Features

🛡️ Core Safety Engine

• Multi-Factor Fatigue Score: Calculates a 0-100% fatigue index based on blink rate, long blinks, and yawn frequency.
• Adaptive Thresholds: Configurable sensitivity settings in settings.json for different driver profiles.
• Thread-Safe Audio: Non-blocking audio playback using pygame and threading for zero-latency alerts.

📡 Connectivity & Services

• SMTP Email Client: Robust email dispatcher with TLS security and error handling.
• Geolocation Fallback: Automatic failover logic for location services.
• Text-to-Speech (TTS): Dynamic voice synthesis for spoken warnings ("Please keep your eyes on the road").

📊 Professional UI/UX

• High-FPS Rendering: Optimized cv2 drawing routines for smooth 30+ FPS performance.
• Visual Debugging: Skeleton tracking overlay for hands and face mesh verification.
• Dark Mode: Reduces screen glare for night driving safety.

---

🛠️ Installation & Setup

Prerequisites

• Python 3.10+
• Webcam (Built-in or USB)

1. Clone & Install

git clone https://github.com/DepthStrider-x/Driver-Safety-Monitor.git
cd Driver_Safety_Dashboard
pip install -r requirements.txt

2. Configure Credentials

1. Open config/settings.json.
2. Update the email section with your SMTP credentials (e.g., Gmail App Password).

   "email": {
       "smtp_server": "smtp.gmail.com",
       "sender_email": "your_email@gmail.com",
       "sender_password": "your_app_password"
   }

3. Add Emergency Contacts

Edit config/contacts.json to define who receives SOS alerts:

"emergency_contacts": [
    {
        "name": "Family Member",
        "email": "receiver@example.com",
        "relation": "Spouse"
    }
]

4. Run the System

python main.py

---

⚙️ Configuration Guide

The system is highly customizable via config/settings.json:

Parameter	Default	Description
sleep_time	2.0	Seconds of eye closure to trigger DROWSINESS alert.
yawn_time	3.0	Seconds of mouth open to trigger YAWN alert.
inattention_time	2.0	Seconds of looking away to trigger FOCUS alert.
sos_hold_time	5.0	Seconds to hold "Namaste" gesture for SOS.
sos_cooldown	300.0	Seconds to wait before allowing another SOS trigger.

---

🔮 Future Roadmap

• Mobile Integration: Companion Android/iOS app for push notifications.
• Cloud Sync: Upload session logs to AWS/Firebase for fleet analytics.
• IR Camera Support: Enhanced night vision using infrared sensors.
• Heart Rate Monitoring: Integration with rPPG (Remote Photoplethysmography) for vital sign tracking.

---

🤝 Contributing

This project demonstrates production-ready code quality, including:

• Strict type hinting and modular design.
• Comprehensive error handling and logging.
• Scalable service-oriented architecture.

Pull requests are welcome. For major changes, please open an issue first to discuss the proposed change.

Developed by Aryan Prajapati