This is our first project for Robotix Club, IIT Hyderabad. This is a small device which tracks 4 physical states.
- Sitting
- Standing
- Walking
- Running
Component Used - Genuino 101 Genuino 101 is a learning and development board based on Intel Curie module. The board recognises gestures, and features a six-axis accelerometer and gyroscope. It also hass a inbuilt Bluetooth Low Energy (BLE). The intel curie module has a inbuilt Pattern Matching Engine. It has a 128 node hardware neural network with two classification alogirthms - the RBF kernel SVM(Support Vector Machine) and K Nearest Neighbours Classification.
- Train the Genuino 101 Board - We collected datasets for various sensor values in 4 physical states. We trained the board using CuriePME library. Code is given here.
- The prediction using RBF kernel SVM was transmitted to android app.
- The android app recorded the values and stored it on our Google Firebase backend.
- Daily stats were made available to users and appropriate recommendations were made.