Inspiration
Many people have the habit of rubbing eyes, scratching, pulling hair etc., subconsciously. We wanted to help people overcome such habits by making them aware.
What it does
It is trained to recognise certain gestures and alert the person.
How we built it
We built a prototype of this device using an Arduino which acquires motion information from an accelerometer and sends it to a PC via Bluetooth. In the PC, we used ESP software which makes use of gesture recognition toolkit and trained the system to recognise the desired gesture. Once trained, if the user performs that gesture, the arduino triggers a buzzer.
Challenges we ran into
We initially tried using signal processing algorithms on the Arduino which was inaccurate. So, we moved to machine learning using ESP on PC.
We had to ensure reliable data over the bluetooth channel by configuring the proper sampling and baud rates.
We had to configure the ESP software and train the system for good performance.
Also, Arduino MKR1000 was a new platform with lack of documentation.
Accomplishments that we're proud of
We are proud that we were able to make a functional prototype in a single day.
What we learned
To explore and solve a challenge with limited time and resources. We are excited to be able to apply machine learning for the first time in our project.
What's next for Aid for subconscious habits
We want to make this feature available to many by implementing it on a wearable like a smart watch which is powerful enough to do ML onboard.
Built With
- arduino
- gesture-recognition

Log in or sign up for Devpost to join the conversation.