When you have attention problems it can be hard to know what exactly works for you when you study, and its also just... hard to sit down and make yourself study! We came up with the idea of an 'accountabilibuddy' for studying that encourages you to do your best while also tracking how well your brain is working so you can figure out exactly what works for you personally while you study and what doesn't! We really hope STUDI-BOT can eventually turn into a system that can actually help people who have attention and learning problems feel encouraged to work hard and do their best!

What it does

Combining the pomodoro method with Pavlovian positive reinforcement and the MUSE headband brainwave tracking we made a system that tracks how well you're focusing over a period of time. The study period involves 25 minutes of work, and 5 minutes of break, on a successful break the robot will dispense a candy to encourage the user! To make the app more user friendly we designed a non-threatening character, "Charles" to keep you going and cheer you on! At the end of your study session, a graph of your brain wave activity over your study period is displayed so you can figure out what worked for you during your study session and what didn't.

How we built it

Using python tkinter we made a gui that shows the image of the character during different phases, working, pausing, and resting (for breaks). Using the existing MUSE jdk to access and record data streamed from MUSE into matlab and we plotted the data using a python script using scipy and matplotlib, to show the brain activity of the user over time. The robot was built using cardboard and plastic bottles we found around the hackathon! We constructed in the robot a lever system that mimics that of old fashioned Amish m&m "machines" but motorized with a microservo, and powered by the arduino. The python tells the arduino when to drop candy through serial, and then the arduino moves the servo to two predefined locations. At the end of the session the graph of brain activity is shown in a different window

Challenges we ran into

We started with doing an android app but discovered we might not be able to communicate from the computer for the arduino to run to the phone, because of limitations on UMBC's wifi network, so we switched to a gui python app. Then muse had some problems with their SDK not being completely available for PC version, so we had to scavenge on the internet for a different solution, eventually, downloading an earlier version of the sdk from google's archives worked. We also had some mechanical problems with converting rotational movement from the servo into horizontal movement for the drawer like system, so he turned out much bigger than we thought he would be and his design varied from what we originally wanted, which was a physical version of the images we made, into a different dude, which is fine because it just means that instead of one friend, you have two! (But in the future we would like him to look like the image!

Accomplishments that we're proud of

The python gui is multithreaded to work more efficiently and the drawer system works with multiple candies in the system (as it should), also because we weren't even sure if we could get data from the muse, we are very happy that we got the muse to give us data :)

What we learned

We learned how to work with python TK in a way that updates the timer every second for countdown, how to use the MUSE's data collection software to get data from the head sensor, and when to give up when the UMBC wifi just sucks at giving a reliable ip.

What's next for STUDI-BOT

We would like for it to have a smaller more efficient physical interface, one that isn't as... clunky (and looks like the original concept), and it would be cool if we could get real time data to alert the user when they aren't paying attention to their task, which would require more time with the muse and a fair amount of data collection and machine learning.

Share this project: