Smart Pill Alert — Connected pill dispenser for remote caregivers
A pill dispenser that reminds the user when to take medication and alerts a remote caregiver if a dose is missed or taken late. Built as a 48-hour pivot at the Ideas Factory competition (University of Málaga, 2019), where it won the Technology Prize at IFSUMMIT 2019. Designed the electronics and control firmware on an ESP32, with a server-side alert path over WiFi.
Problem
Medication adherence is one of the silent failures in elder care. The patient forgets a dose, the caregiver doesn't know until the next visit, and a missed pill compounds with the next one. Phones have alarms but they don't know if the user actually took the pill — they just stop buzzing after the user dismisses them.
Solution
A connected pill dispenser that detects when a compartment is opened, logs the event, and reports it to a small server. If no opening is registered at the scheduled time, the device alerts the user locally and, if needed, sends a notification to the remote caregiver. The prototype is built on an ESP32 with a micro-switch per compartment and a small HTTP server backend.
Key takeaways
- ESP32 + WiFi: a device that wakes on event, posts to a small REST API
- Per-compartment micro-switch detects when the user opens each dose
- Scheduled-dose logic with local buzzer + remote caregiver alert on missed dose
- Built in 48 hours as a competition pivot, then won the Technology Prize at IFSUMMIT 2019
- Breadboard prototype, no custom PCB; firmware in C/C++ on bare-metal ESP32
- Team of six: electronics, 3D modelling, marketing, sales, account management
Where it came from
This project was a competition pivot, not a planned one. In October 2018, the eighth edition of Ideas Factory — a 48-hour entrepreneurship contest at the University of Málaga — asked for a business idea. My team (six of us: electronics, 3D modelling, marketing, sales, account management) spent the first 24 hours on a different idea: a QR-coded wristband that would detect whether a person had suffered a heart attack. The jury pushed back on it and we were running out of time.
In the remaining half of the contest, with the deadline closing in, we pivoted to a connected pill dispenser. The pitch was simple: a device that knows whether the user actually took the medication, and tells a remote caregiver when something goes wrong. We built a breadboard prototype in the second 24 hours. The jury accepted it, and we got the slot for the regional IFSUMMIT 2019 in Seville, where we won the Technology Prize against 17 other university projects.
What I built
My role in the team was the electronics and control. Specifically:
- Schematic and breadboard wiring of the dispenser, on an ESP32 with a micro-switch per pill compartment
- Firmware in C/C++ on bare-metal ESP32: a small scheduler that knows the dose times, listens for the switch events, posts each event to a backend over WiFi
- HTTP backend in a lightweight server: receives the dose events, computes adherence, sends an alert to a caregiver endpoint if a dose is missed
The user side: the dispenser has compartments labelled by dose, the device has a small buzzer that beeps at the scheduled time, the user opens the compartment, the micro-switch closes, the ESP32 wakes from a low-power state, posts the event, and goes back to sleep. If the user does not open the compartment within a grace window, the device escalates: louder beep, then a notification to the caregiver’s endpoint.
What I learned
The interesting lessons from this project were not in the firmware. The interesting lessons were in the pivot:
- A small team under a hard deadline can produce a real working prototype if the scope is small enough. The original idea (heart attack wristband) failed because it was technically ambitious. The pill dispenser worked because every piece of it was within reach: an ESP32, a micro-switch, a WiFi network, a small HTTP server. The constraint forced a better idea.
- Knowing the failure modes of a domain helps a lot. Most of the time in the contest was spent thinking about when a medication gets missed and what a caregiver can do remotely — not on the wiring. The electronic design came almost last.
- The competition gave us something a side project rarely does: a deadline that has to be shipped. Without the 48-hour window, I would not have finished a working prototype. With it, the prototype was on the table by Saturday afternoon.
The project was the prototype + the prize. The team did not start a company. The idea itself — connected adherence for chronic medication — is now a crowded space (smart pill bottles, Medisafe, etc.), which suggests the pain was real and the timing was just slightly too early for a six-person university team to take to market.
What I would do differently today
The electronics were the easy half. Looking back, two things stand out:
- The whole project should have been a real product brief, not a contest pitch. A 48-hour prototype is a great way to validate demand, but a pill dispenser that lives in someone’s home needs regulatory thinking (medical device, GDPR on health data), mechanical thinking (small-batch enclosure for a real pill container), and supply chain thinking (battery life, parts sourcing). We didn’t have any of that, and the prize was a wrong signal: the jury liked the idea, the idea didn’t survive contact with the real product problem.
- The detection method was too coarse. A micro-switch tells you the compartment was opened, not that the pill was actually taken. A load cell under each compartment would tell you that the pill left; a vibration sensor would tell you a hand was there. Either is doable now and wasn’t a stretch for the contest budget.
The hardware is gone, the team went different ways, and the breadboard prototype is in a drawer somewhere. The idea is still a good one — adherence is still broken — and a small batch of connected dispensers would probably find a market today. The right team, a real product person, and a year of runway.