MediAmigo - Technology for well-being
An intelligent assistance system for elderly medication adherence and daily wellbeing.
Motivation
This project was born from a personal experience with my grandmother Luz, who once confided that she sometimes thought she was sick, simply because she forgot to take her medication. As an elderly person who never adapted to modern technology, setting a phone alarm was not a viable solution for her.
MediAmigo was built to support caregivers in ensuring treatment adherence through a personalized, friendly experience designed for elderly users, helping prevent health complications and reduce the risk of hospitalization.
What It Does
MediAmigo allows caregivers to register medication reminders in natural language, which are then converted into audio played through speakers accompanied by blinking LED lights to capture the elderly person's visual attention.
Beyond just medication, the system also focus on preventing health-related issues. MediAmigo generates hydration and clothing alerts based on ambient temperature, also delivered as audio with visual support.
With the press of a single button, the elderly user can send a pre-configured SMS to their primary caregiver with information about their physical and emotional state.
Technology Stack
Hardware
ESP32 microcontroller, LED lights, physical button, sensors, speakers
AI
OpenAI GPT-3.5-turbo for natural language processing: Extracts medication name, time, frequency and duration from plain text input
Azure
Text-to-Speech (es-CR-MariaNeural voice), Blob Storage for audio hosting, Cosmos DB for reminder storage
Backend
Python with Flask. Modules for: LLM handling, TTS generation, Blob Storage export, Cosmos DB and Twilio SMS integration
Impact
MediAmigo bridges the gap between elderly users and modern technology by removing the need for any screen interaction. Its audio-first, visual-aid design makes it accessible to users who cannot adapt to smartphones or digital interfaces.
The Lego-based construction keeps costs low and makes the system replicable by anyone with basic technical knowledge. The project is fully open-source under a Creative Commons CC BY-NC-SA 4.0 license.
Documentation
For technical details on the program flow, visit the repository documentation:
View Repository