Accident Detection System

This project is an enhanced accident detection system with a real-time web and mobile application for alerting and monitoring accidents. It integrates sensor data, machine learning models, a mobile app for on-the-go monitoring, and a frontend dashboard for improved usability and road safety.

Features

• Real-time Accident Detection: Leverages accelerometer and gyroscope data to identify accidents.
• Mobile Application:
  • Developed using Android Studio for on-the-go monitoring.
  • Provides accident alerts and updates on hospital bed availability.
• Web Application:
  • Built with React for visualizing accident data.
  • Features bed availability updates and alert notifications.
• Machine Learning Models: Includes pretrained Dense Neural Network and Isolation Forest models.
• Comprehensive Dataset: Updated and extensive datasets for training and testing.
• Backend Functionality: Scripts for processing data, integrating alerts, and backend operations.

Repository Structure

• App/: Android-based mobile application files.
  • MainActivity.java: Entry point of the application.
  • ui/home/: Home screen of the application.
  • ui/dashboard/: Dashboard for accident monitoring.
  • ui/notifications/: Notification center for updates.
• Trained model/: Contains pretrained machine learning models.
• Phones_accelerometer_reduced/ and Phones_gyroscope_reduced/: Reduced accelerometer and gyroscope datasets.
• gps_dataset/: GPS datasets for training, controlled testing, and uncontrolled testing.
• Test/: Backend scripts for database integration and testing.
• webpage/: React-based frontend application files.
• Python Scripts:
  • integrate_and_post_alert.py: Combines detection results and posts alerts.
  • acquire_bed.py: Estimates hospital bed availability.
  • backend_isolation_only.py: Backend operations using Isolation Forest model.
  • convert.py: Utility for data preprocessing.

Installation

Backend Setup

1. Clone the repository:

   git clone https://github.com/anandharshit712/Accident_Detection_System.git
   cd Accident_Detection_System

2. Install dependencies:

   pip install -r requirements.txt

3. Run backend scripts:

   python integrate_and_post_alert.py

Frontend Setup (Web Application)

1. Navigate to the webpage/ directory:

   cd webpage

2. Install React dependencies:

   npm install

3. Start the React development server:

   npm start

Mobile App Setup

1. Navigate to the App/ directory and open it in Android Studio.

2. Build the project to install dependencies.

3. Run the app on an emulator or physical device.

Usage

1. Run Backend:

   • Execute backend scripts to process sensor data and detect accidents.

2. Use Frontend:

   • Open the React web application to view accident alerts and hospital bed availability.

3. Use Mobile App:

   • Launch the mobile app for on-the-go access to accident alerts and hospital data.

4. Analyze Results:

   • Use pretrained models or retrain with the provided datasets for improved detection.

Datasets

• Training Data:
  • Accelerometer: Phones_accelerometer.csv
  • Gyroscope: Phones_gyroscope.csv
  • GPS: gps_dataset/dataset_training
• Testing Data:
  • Controlled Tests: gps_dataset/dataset_controlled_test
  • Uncontrolled Tests: gps_dataset/dataset_uncontrolled_test

Contributing

Contributions are welcome! Open an issue or submit a pull request for suggestions or improvements.