Case Study: AlignMate Real-Time Posture Correction System

The Challenge: Combating the Modern Epidemic of Poor Posture

As part of a thesis project for the Polytechnic University of the Philippines, our team aimed to create an innovative and engaging solution to the growing problem of poor posture, which is exacerbated by long hours spent in front of computers. The primary challenge was to build a system that could provide accurate, real-time, and personalized feedback without using intrusive cameras for privacy reasons. The solution had to be effective, motivating, and user-friendly, applying many of the universal design principles.

Our Solution: An Integrated Ecosystem for Posture Improvement

I led the software development for AlignMate, a comprehensive system that pairs a custom-built hardware device using an Inertial Measurement Unit (IMU) sensor with a machine learning model and a feature-rich web application. The platform is a complete ecosystem designed not just to correct posture, but to make the process of improving it engaging and sustainable.

• Real-Time Posture Monitoring & Alerts: The system provides a live data stream from the hardware sensor, offering immediate visual feedback on the user's posture. Dynamic notifications alert users the moment they start slouching.
• Intelligent, Personalized Calibration System: A guided, multi-step process allows the user to train a personalized machine learning model. By recording data for their "good" and "bad" posture, the system ensures the feedback is accurate and tailored specifically to them.
• Interactive Data Dashboard: A user-friendly dashboard displays aggregated posture data, daily performance summaries (e.g., "85% good posture today"), and 7-day trends, allowing users to track their progress and identify patterns over time.
• Gamified Achievements & Progression: To keep users motivated, we included gamified elements. Users earn points for maintaining good posture, which can be used to unlock rewards and make progress in themed modules like 'Pet Care' or 'City Builder'.
• In-Depth Analysis & PDF Reporting: The system provides detailed analysis and personalized recommendations. Users can also export their daily reports as well-formatted PDFs for a comprehensive overview of their posture data. The intuitive interface of this feature highlights my focus on delivering a great custom web design with a user-centric approach.

A Multi-Faceted Technology Stack

This project required the integration of hardware, machine learning, a modern web frontend, and a real-time backend to deliver a seamless user experience:

• Frontend: React.js for building a dynamic and interactive user interface, with React Router for seamless navigation between different sections of the application.
• Backend & Real-time Database: Google Firebase was the core of our backend, using its Realtime Database for instantaneous data streaming from the sensor, user authentication to manage profiles, and hosting services.
• Data Visualization: We used the `react-native-chart-kit` library (adapted for web) to create the various charts and graphs on the dashboard that made data easy to understand.
• PDF Generation: The `jsPDF` and `jspdf-autotable` libraries were implemented to allow users to generate and export their detailed data reports.

This thesis project was an incredible learning experience that blended hardware, machine learning, and web development. Interested in integrating IoT or machine learning into your next web project? Get in touch! I also provide API integration services.