Nowadays, there are a lot of disruptive factors in our lives, which are especially disturbing when it comes to studying or working, when a high level of concentration is required. During these cognitively demanding sessions, it is very hard for one to objectivly identify these factors and exclude them from his environment. Once we become aware of our bad habits, it is much easier to fix them therefore making our studying more effective.Storyboard
Our solution uses an algorithm that identifies disruptive sounds during your studying sessions. It specifically searches for sounds of notifications, speech, and everything else that might be distruptive noises in your environment. At the end of your session, our system informs about the patterns detected. Based on the recieved information, you can fix these bad habits in your next session.Description
The system consists of an Arduino Nano 33 BLE Sense and an Android application. Firstly, the user starts the monitoring using our Android app. When the session is started, the app collects data about the session through BLE. What's more, the Arduino board LED changes color when there are disruptive sounds in the users environment, therefore warning him of the disturbing factors in real time. Finally, when the user ends the session (using the app), the app shows statistics of it.Hardware & Connectivity
For this project we decided to use the Arduino Nano 33 BLE Sense. To connect to our Android app, we used the boards BLE support. For sound monitoring we used the built-in michrophone MP34DT05.Machine Learning
Our machine learning model has three different classes: Effective, notifications and speech. Accuracy of our whole model is around 94%.
The android app is written in java. We use Android's BLE support to recieve the data transmitted by the Arduino board. For the statistics pie chart, the liprary MPAndroidChart is used. When the user opens the app, it automatically connects to the Arduino board, by searching for its BLE name.
We believe our prototype represents a good starting point to evolve our solution and create a more advanced and user-friendly system. In the future, we would make the following improvements:
- Expand the ML model to detect a wider range of disruptive noises
- Implement the viewing of sessions' histroy in our app together with some advanced statistics
- Achieve an even better user experience