CUBoulder-2018spring-ML4HCI / G-Acc-Perez-Fermin-Joshi

We put the repository in the wrong spot - so we're resubmitting it in the correct spot today

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

#Micro-Bit PY Game

Developers

Jonathan Fermin, Robert Perez, Kushal Joshi

Goals

To create a working PY game Using MicroBit Controlling the PYGame with MicroBit.

Tools

Microbit

Processing

Wekinator

Pygame

Approach

We used the MicroBit movement and accelerometer data to move a virtual blob in a Pygame.

Sensors used

Everything is done through the MicroBit

Feature extractor approach

Our Wekinator setup dealt with the 3 input, 1 output (4 classes). The 3 inputs are based off accelerometer data provided us with the 3 input data. Our output was the 1 output with 4 classes which represented the 4 dimensions.

ML model structure - what you did and why, including results from experiments you tried along the way

We used K nearest neighbors, which was effective. The KNN approach was proved to be simplistic, which was perfect for our 2-Dimension. We only had to worry about the x,y direction but if we had the 3rd dimension we might have needed to explore other ML models.

Ideas about how other might improve upon your work

  • Implementing this to a 3-D space would be our next goal in mind. Being able to move in a 2-D space is very neat, but provides limited real world application.

*Using the 2-D space that we have to work with, implementing this to a wheelchair setting would be a great topic to leap off from. Allowing a wheelchair user to use a device to tilt to move themselves.

Demo Video

Demo Video Link

About

We put the repository in the wrong spot - so we're resubmitting it in the correct spot today


Languages

Language:Processing 85.2%Language:Python 14.8%