🚀 NASA Sentinel | Edge AI Predictive Maintenance

Sentinel is a high-performance Edge AI inference system designed for real-time predictive maintenance of NASA turbofan engines. Built with an ultra-lightweight C++ core and a stunning glassmorphism-styled dashboard, it provides instantaneous health diagnostics for complex aerospace systems.

---

✨ Key Features

• ⚡ Ultra-Fast C++ Engine: A custom inference engine built for speed and efficiency, capable of running on low-power edge devices.
• 📊 Real-Time Dashboard: A premium, interactive web interface for monitoring fleet health and simulating sensor failure states.
• 📡 RESTful API: Seamlessly integrate with external sensors via the /predict endpoint.
• 🐳 Containerized Deployment: Multi-stage Docker builds ensure a tiny footprint and "run anywhere" capability.
• 🛠️ Diagnostic Simulator: Full 17-axis sensor override using real NASA CMAPSS data ranges.

---

🏗️ Architecture

The system is split into three primary layers:

1. The Brain (sentinel_model.bin): Quantized weights and inference logic optimized for edge performance.
2. The Engine (compiler/main.cpp): A C++ service that loads the model, provides the REST API, and serves the static dashboard.
3. The Interface (web/): A modern, responsive dashboard built with vanilla CSS and JS, utilizing glassmorphism for a premium look and feel.

---

🚀 Quick Start

Running with Docker (Recommended)

The fastest way to get Sentinel up and running is via Docker:

# Build the production image
docker build -t sentinel-edge-ai .

# Run the engine
docker run -p 8080:8080 sentinel-edge-ai

Once running, navigate to http://localhost:8080 to access the dashboard.

Local Development

To compile and run locally, ensure you have a C++ compiler and the necessary headers:

1. Compile the engine:

   g++ -O3 -o sentinel_engine compiler/main.cpp

2. Start the server:

   ./sentinel_engine

---

🛠️ API Reference

POST /predict

Submit sensor data for real-time analysis.

Payload: A comma-separated string of 17 sensor values. Example:

curl -X POST -d "642.68,1590,1409,553.4,2388.1,9065,47.5,521.4,2388.1,8144,8.44,393,38.8,23.29,..." http://localhost:8080/predict

Response:

{
  "probability": 0.012,
  "status": "NOMINAL"
}

---

📁 Repository Structure

.
├── compiler/           # C++ Inference Engine source code
├── web/                # Glassmorphism Dashboard (HTML/CSS/JS)
├── data/               # NASA CMAPSS dataset samples
├── kernels/            # (Future) Optimized CUDA kernels
├── Dockerfile          # Multi-stage production build script
├── sentinel_model.bin  # Compiled model weights
└── README.md           # You are here

---

🛡️ License

This project is for educational and research purposes, utilizing publicly available NASA datasets.

---

Developed for the Edge of Tomorrow — "Silicon to Synapse"