🛰️ Debris Guardian – AI Space Traffic Management System

Debris Guardian is a full-stack, AI-driven Space Traffic Management system designed to mitigate the growing threat of satellite collisions in Low Earth Orbit (LEO). The project simulates real-world orbital dynamics and leverages advanced machine learning techniques to predict, prevent, and manage space debris risks.

---

🚀 Overview

With the increasing number of satellites in orbit, collision risks are rising significantly. This project addresses the problem by combining:

• Reinforcement Learning for autonomous collision avoidance
• Predictive Analytics for untracked debris risk estimation
• Multi-Agent Systems for distributed decision-making
• Real-time simulation and visualization

---

🧠 Key Features

🔹 1. Reinforcement Learning (PPO)

• Implemented a Proximal Policy Optimization (PPO) agent
• Custom simulation environment using Gymnasium
• Learns optimal satellite maneuvers to:
  • Avoid collisions
  • Minimize fuel consumption

---

🔹 2. Predictive Modeling (Random Forest)

• Built a Random Forest model to estimate risks from small, untracked debris (1–10 cm)
• Uses historical fragmentation data to simulate real-world conditions

---

🔹 3. Multi-Agent System (MAS)

• Designed a distributed AI architecture with specialized agents:
  • Lookout Agent → Detects threats
  • Diplomat Agent → Handles negotiation between satellites
  • Pilot Agent → Executes optimal maneuvers

---

🔹 4. Orbital Mechanics Integration

• Uses real-world satellite propagation models (SGP4)
• Processes TLE (Two-Line Element) data for realistic satellite positioning
• Simulates accurate orbital movement and collision scenarios

---

🔹 5. Satellite-to-Satellite Negotiation

• Implements Contract-Net Protocol
• Enables autonomous decision-making between satellites on collision paths
• Optimizes which satellite should maneuver

---

🔹 6. Delta-V Optimization

• Calculates minimal thrust required for orbit adjustments
• Ensures fuel-efficient satellite movement

---

🔹 7. Real-Time System

• Backend built with FastAPI
• Real-time communication via WebSockets
• Streams telemetry data continuously

---

🔹 8. 3D Visualization Dashboard

• Built using React and Three.js
• Displays:
  • Satellite positions
  • Debris fields
  • Risk alerts
  • Live system telemetry

---

🏗️ Tech Stack

💻 Backend

• Python
• FastAPI
• WebSockets
• Gymnasium

🤖 Machine Learning

• Reinforcement Learning (PPO)
• Scikit-learn (Random Forest)
• NumPy, Pandas

🌐 Frontend

• React.js
• Three.js

⚙️ Tools

• Git & GitHub
• Docker (optional deployment)

---

📊 System Architecture

        +----------------------+
        |   Frontend (React)   |
        |  Three.js Dashboard  |
        +----------+-----------+
                   |
            WebSockets (Real-time)
                   |
        +----------v-----------+
        |     FastAPI Server   |
        |  Simulation Engine   |
        +----------+-----------+
                   |
+------------------+------------------+
|                                     |
+-------v--------+ +--------v--------+
| RL Agent (PPO) | | Random Forest |
| Collision Avoid| | Risk Prediction |
+----------------+ +-----------------+
|
+-------v--------+
| Multi-Agent Sys|
| Lookout/Diplomat|
| Pilot Agents |
+----------------+

---

⚙️ Installation & Setup

1️⃣ Clone the Repository

git clone https://github.com/SharathWeb/debris-guardian.git
cd debris-guardian

### 2️⃣ Backend Setup
pip install -r requirements.txt
uvicorn main:app --reload

### 3️⃣ Frontend Setup
```bash
cd client
npm install
npm run dev

---

🌐 Deployment (Hybrid: Vercel + Render)

This project is configured for a hybrid deployment to support real-time WebSockets.

🔹 1. Backend (Render)

1. Create a new Blueprint project on Render.
2. Connect this repository.
3. Render will use render.yaml to automatically:
   • Set the root directory to server/.
   • Install dependencies from server/requirements.txt.
   • Start the FastAPI server using Uvicorn.
4. Important: Add your .env variables (Snowflake, API Keys) in the Render Dashboard under Environment.

🔹 2. Frontend (Vercel)

1. Import this repository into Vercel.
2. Vercel will detect the vercel.json in the root and build the client/ directory.
3. Environment Variables: You must add these in the Vercel Dashboard:
   • VITE_API_URL: Your Render backend URL (e.g., https://your-app.onrender.com)
   • VITE_WS_URL: Your Render WebSocket URL (e.g., wss://your-app.onrender.com/ws)

---

▶️ Usage • Run backend server • Start frontend • Open dashboard in browser • Observe: • Real-time satellite movements • Collision alerts • AI decision-making

📌 Applications • Space Traffic Management • Satellite Collision Avoidance • Aerospace AI Systems • Autonomous Decision Systems

🔮 Future Improvements • Integration with real satellite APIs • Advanced deep learning models for debris prediction • Scalable cloud deployment (AWS / Azure) • Enhanced visualization and analytics

👨‍💻 Author Sharath S Prasad

📧 sharathspd1818@gmail.com 🔗 LinkedIn: https://www.linkedin.com/in/sharath-s-prasad-204a28227