This week has been intensive in terms of work and study. During the course of my second week at the neuromoph workshop in Telluride, CO, I have joined a project group that involves both game design and neuromorphic recordings from the brain. Here’s the project:
The project is a mountain biking game that aims to record both sensory error feedback and motor brain activity. The game accomplishes to collect these two specific brain functions by the player using a mouse (motor) and the player seeing the error that they have achieved (sensory). The game is basic: stay on a trail and how ever much the player deviates from that trail they will see a red error bar indicating their error. This game is a continuation from last year, but the two directors of it have essentially given me full authority of making the game look and play better and differently. The ways I changed up the game are by making it look better visually (easy) and reworking a lot of the basic game mechanics (ehh). In the end, the data collected from the game should be adequate enough to compare sensory and motor responses, and hopefully provide a basis of isolating both brain functions into their own individual components. This separation of sensory and motor can allow interesting ideas to arise; like playing the game while another computer makes the move 1 second before you do by interpreting your motor response. Me and the other people on the team hope to come up with ideas to achieve this next week. Now onto the changes I’ve made to the game:
The game used to look like this ->
But now it looks like this (wow) -> 
The visual part was the least of my concerns though.
In terms of rework, the initial game went on for 45s and recorded your error. Now, the game has 10 different difficulties, of 60s in length, that record the error and then play back the game play. I took care of making the levels while one of the other project group members worked on the playback. The levels were honestly easy because I just set a seed for the random generation of the trail and then based on that seed had different difficulties. I had to do a lot debugging though because the code didn’t incorporate any organization like OOP.
Secondly, I had to make a brand new demo game mode that was going to be shown at a demo my team had to do on Friday. So essentially I had two days to do both demo and level rework. The original game I guess had a demo but it was simply not for collecting data. The demo I was told to make incorporated an increase in trail size, infinite terrain, and finite error to make. After a day of working on the game with an updated leaderboard system it was done.
The game works by shrinking the trail as time goes on, and when you touch the green, it deducts error and auto plays for a bit to get you back on track. A lot of this coding has been done in python using the pygame library which made this project easy (if your not taking into account how the code was developed).
Besides designing the game, I went to many lectures too, that ranged from a multitude of different topics.
Next week I hope to again use the EEG (electrode brain recorder) to then collect data and move on from there.
In terms of what I’ve done outside of the neuromorph workshop, I have been continuing my web dev hobby, trying to make a neural net, and tubing (the water was cold).
There are no comments published yet.