During this global crisis, a large proportion of the elderly and nursing home bound population is at a greater risk of fatalities and exposure to covid19. Also, staying locked down and under quarantine, we become neglectful of our health. We built a solution to address these issues and more, using a combination of hardware and software to make the next generation healthcare wearable device.
What it does
The hardware device is wireless and battery-powered and can detect the following (sensor name in parentheses): Heart rate /pulse (MAX30102) Blood oxygen saturation level (MAX30102) Fall detection (6-axis IMU) Number of steps taken (6-axis IMU) Body temperature (DS18b20 waterproof contact temperature probe) Galvanic Skin Response (self-built hardware with resistors, tape and aluminum foil) Ambient (outside) temperature (DHT 12) Ambient Humidity (DHT 12) Ambient air pressure (BME 280)
This is coupled with a mobile app that allows us to view these readings in real-time as well as view historical data for tracking and analysis. The fall detector is connected to an emergency call service which will alert emergency services unless the alert is canceled.
Finally, the wearable also has a built-in UV LED sanitizer for sanitizing surfaces like door handles before using them.
The app shows …..
How we built it
The hardware base of this project is an M5Stick C microcontroller. M5Stick C is powered by ESP32 microcontroller and has a built-in 6-axis IMU. The IMU takes the accelerometer and gyroscope reading which are then used to calculate the number of steps. External hardware plugins also combined with the M5 StickC which provides more information discussed above.
The hardware comes with an open-source app which takes the reading from all the sensors and displays it to the user in an easy to use interface.
The hardware is esp32 based so we wrote a connection routine that connects to the local wifi network and hits the backend which is implemented as APIs running on a virtual server on GCP for the most part and a local server on our pc for the fall detection alerting mechanism. The data is all streamed to a database on MongoDB and the backend also implements querying functionality for the app to gather this data.
The mobile app was built using React Native and Expo for quick cross-platform development. React-native-chart-kit was used for visualizing the data.
Challenges we ran into
M5 Stick C is a fairly new hardware and there is not a lot of online support for troubleshooting issues. Moreover, since the device was originally made in china, the translations and explanations of some major functions were not clear.
Accomplishments that we're proud of
Working, wearable device with real-time data Functioning backend servers for data management Easy to use app which can present real-time data to the user
What we learned
Learned a lot about ESP32 and M5 Stick C Troubleshooting using Chinese forum and google translate Importance of galvanic skin response for detecting moods and emotions.
What's next for Get Me Fit
In terms of enhancing the product, we can identify the user’s state of mind using GSR readings and machine learning. In terms of marketing the product, we would like to initially target nursing homes and the elderly.