We tried to detect dangerous situations through smartwatch. We created a service that detects two potentially dangerous situations through the IMU(Inertial Measurement Unit) sensor present in all smartwatches and sends them to the phone to alert the user. To show the feasibility of the service, we tested it using Arduino Nano 33 BLE Sense, and through this, the accuracy of voice and gesture recognition was measured to be 92 or 93%, implying that it will perform well when applied to real-world scenarios.
Libraries and frameworks
- Classifier: Google colab
- Emergency detection algorithm: arduino
- Communication with BLE: Python