We wanted to make a low cost medical alert wearable that works on WiFi and Tropo. There are similar products on the market, but they are prohibitively expensive and impractical for low income households. We aim to create a product that can serve the need at significantly lower price points.

Problems with existing products

Medical alert devices already exist in many forms, but the leading solutions are prohibitively expensive both in terms of initial cost and ongoing cost. This forces out one of the largest potential customer bases for these products: low income families. In addition, many of these solutions have not been updated since their inception make inefficient use of modern technology.

How we fixed these problems

With WiFi technology becoming more and more ubiquitous, we decided to use it as a basis for our project's connectivity. And, instead of having a manned call center, we reduced cost by using the Cisco Tropo API to alert a predetermined emergency contact list. Our device is capable of detecting falls through an accelerometer, and a user can manually trigger an alert using an alert button.

Our device runs on a Particle Photon and sends alert data to a flask server. In the case of an alert, the server retrieves a list of the user's emergency contacts and sends a request to Tropo to notify the contacts. The watch uses a capacitive sensor to know if it is on a user's wrist, and displays information to the user via an LED ring. Alerts can be manually triggered using an alert button or detected using an accelerometer. In addition, our watch tells the time.

A detailed description of our state machine is available on the github page.

Challenges we ran into

Our backend stack consisted of multiple different layers and utilized several external APIs. Getting everything to work together and run smoothly took some time, but in the end all of the information was managed appropriately. In addition, the Particle Photon does not work well with many Arduino libraries; interfacing the development board with our hardware was somewhat of a challenge.

What's next for Watch in Motion

Going from prototype to a functional consumer model and implementing advanced functionality such as heart rate monitors or accurate location data.

Share this project: