🏋️‍♂️ Exercise Recognition App

A React Native app that uses motion sensor data to recognize physical exercises using a trained machine learning model. It allows users to start and stop workout sessions, processes the collected data, predicts the exercise being performed, and stores all session information in Firebase.

---

📚 Table of Contents

• About the Project
• Features
• Architecture
• Tech Stack
• Getting Started

---

🧠 About the Project

This mobile application aims to provide a smooth experience for users who want to track their workouts automatically. By collecting sensor data from mobile devices, the app predicts the type of exercise being performed using a hybrid 1D CNN-LSTM model trained on the PAMAP2 dataset (for now).

---

✨ Features

• Start/Stop exercise sessions
• Live prediction of performed exercise
• Firebase Authentication for user management
• Cloud-based ML inference with Firebase Functions
• Logs exercise history and statistics
• Data stored securely in Firestore

---

🏗 Architecture

The system is composed of a React Native frontend and a Firebase backend that handles authentication, model inference, and data storage.

Click to view architecture description

📱 Frontend – React Native App

• Allows the user to start and stop an exercise session.
• Collects sensor data.
• Sends sensor data to the backend for prediction.
• Displays the predicted exercise to the user.
• Manages login and authentication.

🔧 Backend – Firebase

🔐 Firebase Authentication

• Manages user credentials and provides auth tokens.

⚙️ Firebase Cloud Run - Deployed API

• Loads the ML model.
• Predicts the exercise from sensor data.
• Sends prediction to app and stores result.

🧠 Firebase Cloud Storage

• Stores the trained machine learning model.

🗃 Firestore Database

• Stores:
  • Exercise predictions
  • User metadata
  • Session logs
  • Badge information
  • User statistics
  • Feedback from users

---

🛠 Tech Stack

• React Native (Expo)
• Firebase (Authentication, Firestore, Cloud Storage)
• Python for model training
• PyTorch for the ML model
• mHealth dataset for training

---

🚀 Getting Started

Prerequisites

• Node.js and npm
• Expo CLI (npm install -g expo-cli)
• Firebase project with:
  • Authentication enabled
  • Firestore and Storage configured

Setup

1. Clone the repository:

   git clone https://github.com/JoaoCoelho2003/SA-FitSense.git
   cd SA-FitSense

2. Go to the frontend folder:

   cd frontend

3. Install dependencies:

    npm install

4. Set up your Firebase config in the app.

5. Start the app:

   expo start