Micro-Doppler Classification System

Invisible Eye is a cross-platform application built with React.js for web and React Native for mobile, enabling seamless monitoring and surveillance. The app allows users to view real-time data, track activity, and manage alerts across devices with efficient performance. It leverages React's component-based design to ensure a responsive and engaging user experience, and its integration with backend services supports robust data handling and secure operations.

Table of Contents

• Project Title
• Table of Contents
• Deployed Website Link
• Technologies Used
• Installation
  • Backend Setup
  • Web Application Setup
  • Mobile Application Setup
• Backend
• Web Application
• Mobile Application
• Analysis of Feasibility
• Potential Challenges
• Overcoming Strategies
• Contributing
• License

Deployed Website Link

Access the deployed web application at birddroneclassifier.netlify.app

Technologies Used

This project leverages a diverse set of technologies to build a complete micro-Doppler classification system:

• FMCW Radar: Used to capture micro-Doppler signatures for classification.
• Python: Programming language used for data processing and backend development.
• PyTorch: Deep learning framework utilized to build and train the machine learning model.
• SciPy & NumPy: Libraries for mathematical computations and signal processing.
• React.js: Frontend framework used to build the web application.
• React Native: Framework for building cross-platform mobile applications.

Screenshots

• Website

• Mobile Application

Installation

Follow the instructions below to set up each component of the system:

Backend Setup

1. Clone the repository:

   git clone https://github.com/bishwatkur/SIH24.git

2. Navigate to the backend directory:

   cd SIH24/backend

3. Create a virtual environment and activate it:

   python -m venv venv
   source venv/bin/activate   # On Windows: venv\Scripts\activate

4. Install the required dependencies:

   pip install -r requirements.txt

5. Start the backend service:

   python app.py

   The backend will start running on http://localhost:5000.

Web Application Setup

1. Navigate to the web application directory:

   cd ../web

2. Install dependencies:

   npm install

3. Start the web application:

   npm start

   The web application will be available at http://localhost:3000.

Mobile Application Setup

1. Navigate to the mobile application directory:

   cd ../mobile

2. Install dependencies:

   npm install

3. Start the mobile application (emulator or device required):

   npx react-native run-android  # For Android
   npx react-native run-ios      # For iOS

   Ensure that your mobile device or emulator is properly set up.

Backend

The backend service is developed in Python and is responsible for:

1. Data Preprocessing: Converting FMCW radar data into a format suitable for ML model input.
2. Model Deployment: The deployed ML model, built using PyTorch, classifies the micro-Doppler signatures to detect different activities or objects.
3. API Endpoints: RESTful APIs are provided for the web and mobile applications to send data and receive predictions.

ML Model

The model is designed to classify micro-Doppler signatures into predefined categories. It utilizes a convolutional neural network (CNN) architecture optimized for time-series data.

Web Application

The web application serves as an interface for users to upload radar data files and visualize classification results. It provides the following functionalities:

• Uploading micro-Doppler data in compatible formats.
• Viewing classification results in real-time.
• Interactive graphs and visualizations of the radar data and model predictions.

Mobile Application

The mobile application is developed using React Native to support both iOS and Android platforms. It allows users to:

• Capture micro-Doppler data through connected devices.
• View real-time classification results.
• Access classification history and performance analytics.

Analysis of Feasibility

• Scalability: The system is scalable with modular components for backend, web, and mobile applications, allowing independent scaling of services.
• Ease of Setup: The modular structure ensures ease of setup and maintenance.
• User-Friendliness: Both web and mobile interfaces are designed with usability in mind, offering a seamless experience across platforms.

Potential Challenges

• Backend Integration: Ensuring smooth communication between the web and mobile applications with the backend service.
• Cross-Platform Consistency: Maintaining a consistent user experience across web and mobile platforms.
• Real-Time Data Processing: Handling real-time radar data and providing instantaneous classification results.

Overcoming Strategies

• Backend Optimization: Optimize API responses and use efficient data handling techniques to reduce latency.
• Cross-Platform Libraries: Use shared libraries and components for web and mobile applications to maintain consistency.
• Data Management: Implement a robust data pipeline to handle and preprocess large volumes of radar data efficiently.

Contributing

For changes, please open an issue first to discuss what you would like to change.