Edge AI Stability Detection System

A machine learning system for detecting system stability using simulated IMU sensor data.
The project demonstrates an end-to-end ML workflow including data generation, model training, real-time inference, and API deployment.

---

Project Structure

stability-ai-system
│
├── data
│   └── synthetic_sensor_data.csv
│
├── training
│   └── train_model.py
│
├── model
│   └── stability_model.joblib
│
├── api
│   └── main.py
│
├── inference
│   └── realtime_predict.py
│
├── requirements.txt
└── README.md

---

System Pipeline

Synthetic Sensor Data
        ↓
Data Processing
        ↓
Model Training (RandomForest)
        ↓
Model Evaluation
        ↓
Model Persistence (joblib)
        ↓
Real-Time Prediction Simulation
        ↓
REST API Deployment (FastAPI)

---

Features

• Synthetic IMU-like sensor dataset generation
• RandomForest classifier for stability detection
• Real-time prediction simulation
• Model persistence using joblib
• REST API deployment using FastAPI

---

Dataset

The dataset simulates IMU sensor readings including:

• tilt_x
• tilt_y
• angular_velocity

A rule-based labeling system determines whether the system is stable or unstable.

Example:

tilt_x	tilt_y	angular_velocity	stable
1.2	-0.5	0.8	1
8.3	-6.1	4.2	0

---

Installation

Clone the repository:

git clone https://github.com/yourusername/stability-ai-system.git
cd stability-ai-system

Install dependencies:

pip install -r requirements.txt

---

Train the Model

python training/train_model.py

This will:

• Generate the dataset
• Train the model
• Save the trained model to model/stability_model.joblib

---

Real-Time Prediction Simulation

Run:

python inference/realtime_predict.py

Example output:

TiltX:1.20 TiltY:-0.40 AV:0.80 -> STABLE
TiltX:7.50 TiltY:6.20 AV:3.90 -> UNSTABLE

---

API Deployment

Run the FastAPI server:

uvicorn api.main:app --reload

Open the API documentation:

http://127.0.0.1:8000/docs

---

API Example

POST /predict

Input:

{
 "tilt_x": 1.2,
 "tilt_y": -0.3,
 "angular_velocity": 0.7
}

Output:

{
 "prediction": "stable"
}

---

System Pipeline

Future Improvements

• Integration with real IMU sensors (MPU6050)
• Deep learning time-series models
• Edge deployment on embedded devices
• Logging and monitoring system

---

Author

Aaradhya Dev Tamrakar
Electronics, Communication & Information Engineering
IOE – Tribhuvan University