Neuro-Edge-Sentry: AI-Powered IoT Orchestrator

🚀 Overview

Neuro-Edge-Sentry is an advanced predictive maintenance orchestrator designed for Industry 4.0. It solves the critical problem of Resource Contention on Edge Devices (like Arduino/Raspberry Pi) by intelligently scheduling sensor tasks.

Unlike traditional schedulers that react to overload, Neuro-Edge-Sentry predicts it using a Machine Learning Decision Tree trained via Imitation Learning.

🧠 The AI Upgrade

Originally, this project used a heuristic Look-ahead Search (O(b^d) complexity). While accurate, it was too slow for real-time edge deployment.

We cracked this problem using Imitation Learning:

1. Teacher: A rigorous Look-ahead Search algorithm simulated 2,000+ factory scenarios.
2. Student: A DecisionTreeClassifier was trained to "clone" the teacher's behavior.
3. Result: We achieved O(1) inference latency with 99% safety compliance—bringing heavyweight logic to lightweight chips.

✨ Features

• 3-Mode Scheduling:
  • Round Robin (Base)
  • Strict Priority (Naive)
  • ML Decision Tree (Advanced AI)
• Real-Time Dashboard: Streamlit-based UI to visualize CPU/Mem load and decisions.

• Crisis Protocol: Automatically sheds non-critical tasks (e.g., Timelapse) to save critical ones (e.g., Pressure Valve Check).
• Explainable AI: The system logs why it made a decision (e.g., "Cost to IDLE is cheaper than risking Overload").

🛠️ Project Structure

Neuro-Edge-Sentry/
│── dashboard.py            # Streamlit UI with Scheduler Selection
│── final_scheduler.py      # Core Logic (Sim + ML Inference)
│── generate_dataset.py     # [NEW] Data Generator for Imitation Learning
│── train_model.py          # [NEW] Model Training Script
│── decision_tree_model.pkl # [NEW] Trained AI Model
│── scheduler_data.csv      # [NEW] Training Dataset
│── sensor_log.csv          # Sample Sensor Data
│── README.md               # Documentation

⚡ Quick Start

1. Clone & Install

git clone https://github.com/saimayithri/Neuro-Edge-Sentry.git
cd Neuro-Edge-Sentry
pip install -r requirements.txt

2. Run the AI Dashboard

streamlit run dashboard.py

3. Retrain the Brain (Optional)

Warning: This runs 2,000 simulations.

python generate_dataset.py  # Create Data
python train_model.py       # Train Model

---