Developing an IoT based fall detection system and which also sends an emergency SMS alert. The device should alsorecord pulse rate which can be stored and monitored through a web dashboard and trigger an alert on abnormal patterndetection.
We have used two sensors, GY521 (MPU6050) accelerator plus gyroscope and KY039 heart rate sensor. The data fromthe sensors are collected via two separate esp8266 (NodeMcus). The NodeMcu(s) then uses the MQTT protocol forcommunication with the raspberry pi (Gateway) which continuously receive the data, and transmits it to IoT severperiodically.
We set up the IoT server on Azure IoT Hub. We configured a script on the server which saves the received data to aSQLite database. The script also checks for abnormal patterns and triggers an SMS alert. We have used the Twilio APIfor sending alert messages with pulse rate details and timestamps included. We have developed a flask application whichdisplays the fall detection status, device live status and pulse rate history to the user on an intuitive dashboard. Themonitoring application is deployed on pythonanywhere.
- Scalable (can add additional sensors without modifying the existing sensors).
- Edge processing (all the data is processed at the edge nodes before transmission to the cloud).
- Visualization (the data is visualized in the form of a web dashboard).
- Robust (the system does not falter to any false detection because of the trigger based algorithm).
- Real time alert (concerned people are notified in real time about any unusual activity).
The final application is deployed here
Developed By : Sanyam Jain and Dhruv Jain