AeroVision

AeroVision

A real-time flight radar mobile application with AI insights.

Architecture

• Frontend: React Native (Expo) with TypeScript
• Backend: Python FastAPI with WebSockets
• Database: PostgreSQL with PostGIS
• Cache: Redis
• ML: Custom models for delay prediction, turbulence estimation, and anomaly detection

Setup

Backend

1. Install dependencies:

   cd backend
   pip install -r requirements.txt
   

2. Set up PostgreSQL and Redis.

3. Run the backend:

   uvicorn main:app --reload
   

4. Run ADS-B ingestion:

   python ingest/adsb_stream.py
   

Frontend

1. Install Expo CLI.

2. Install dependencies:

   npm install
   

3. Start the app:

   npm start
   

API Endpoints

• GET /flights/live - Live flights
• WS /ws/flights - WebSocket for real-time updates
• GET /flights/{icao} - Flight details
• GET /ai/delay/{flight_id} - Delay prediction
• GET /ai/turbulence/{flight_id} - Turbulence prediction

Sample WebSocket Payload

[
  {
    "icao": "ABC123",
    "callsign": "UAL123",
    "lat": 37.7749,
    "lon": -122.4194,
    "altitude": 35000,
    "speed": 500,
    "heading": 90
  }
]

Sample ML Outputs

Delay Prediction:

{
  "delay_probability": 0.78,
  "reason": "Airport congestion"
}

Turbulence:

{
  "turbulence_level": "Moderate",
  "confidence": 0.85
}