What happens when an elderly person falls and is left unconscious in their own home. Tragically, too many senior citizens are found dead in their homes every year because they fell and had no way to call for help. We were made aware of these stories and the anxiety that some elderly people have living alone through a meeting with JFCS - a local organization that empowers older adults to live independently.
Rescue Me is a wearable device that solves this problem. It's worn on the wrist, and can detect a fall. If a fall is detected, the device vibrates for 15 seconds, awaiting response. If the user pressed the button, that means he or she is ok, and no help is needed, but if there is no response, Rescue Me calls for help. Furthermore, Rescue Me doesn't just call 911, it can be programmed to call the best point of contact, building security, a neighbor, a nurse etc. In addition, we've built a dashboard for nursing homes that can show fall data and frequency to track how often people are falling - this can facilitate recommendations of canes or walkers.
How we built it
To build a prototype of our device, we used a NodeMCU, accelerometer, vibrating motor, and button - all inside of a 3d printed housing. We started out by building the device on a breadboard. This was so that we could iterate quickly and fix wiring errors easily. After implementing fall detection and the user experience logic on the breadboard, we soldered new components onto a prototyping circuit board and fit it into a 3D printed case modeled in Solid Works. We used new components and kept the original breadboard prototype in-tact for reference and in case something went wrong with the soldered version. Our team used Adafruit dashboard to make a display of our user data and alerts associated with calls for help. Using Adafruit dashboard was one of the most challenging parts of the project, and we got banned a few times for sending too much data by mistake. In the end, we refined our logic and only sent data when we absolutely needed to - saving device power and ensuring that we weren't banned.
After finishing up the dashboard's look and feel, we moved onto testing. We coded in many edge cases to ensure that the device wouldn't think that walking around, sitting down in a chair, or writing were falls, for example. Additionally, we dropped the device many times to ensure that it would pick up real falls. We had to make sure that the device was reliable, but that it also didn't have too many false positives - this was a balancing act, for sure.