Closing an Amazing Summer

My time here in Wyoming is sadly drawing to a close.

It has been only just over 10 weeks and yet I feel as though I have learned so much. I was truly blessed with the fortune of helping at the 3DIA lab under great leadership and alongside an amazing team. I had entered nervous and excited at the prospect of research with older students. I was warmly welcomed and made part of the team from the very start. Seeing my modeling skills come to good use early was wonderful, as was adapting my training in other technology to finish experiment, complete the code of an experiment, and designing a new experiment.

The DREU experince was fantastic and I will certainly rave about it to everyone back at my current home instituion, The College of William and Mary.

My Very Own Experiment

After much brainstorming and sleuthing through research papers, I finally found a research question that really drives me. With the help of my professor, we were able to write an Institution Review Board request to experiment.

This was the first IRB proposal that I had written, and while extremely tedious, I found that it is necessary to have a human system in place to protect experiment participants and to stop unethical experiments from starting at research institutions. I was able to learn how to treat participants through the readings the IRB website gave me (alongside the tutelage of Professor Banic). These new interpersonal skills I learned led me to be a better researcher.

While I cannot yet fully disclose my research, I will say that it involves electroencephalography, a tool I learned to use here in Wyoming to gather input from the brain (with the aid of signal processing). I fear I will not be able to run participants here in Wyoming due to my time drawing to a close here but I will do my best to complete all of the code and tools necessary to conduct a preliminary version of the experiment before I leave.

CAVE 

I was recently to see the University of Wyoming’s new CAVE system installation (you can read more about it here ). The expereince was absolutely breathtaking. I often find myself ststrugglingto explain the feeling of using the CAVE system to navigate a rendered town. It’s use of optical tracking is stunning as you are able to see all the walls around you shift to the perspective of your glasses’ current coordinates. Peaking around corners in a pre-rendered environment felt so real that I found myself avoiding them for fear of bumping into them.

After a bit of work, my co-researchers were able to put our old model of the three dimensional dungeon with a bridge into the CAVE system. It was such a delight to truly feel surrounded by he walls I had previously created in Blender3D and just barely being able to see through the cave’s windows. The experience was quite enjoyable but I was left with a lot of questions as to how to improve the system.

The CAVE uses two dimensional menus for adjusting options which feels unnatural and forced in the otherwise amazing system. It also is difficult typing within the system as you must be standing up when using it. I believe a new method of interacting with the system and a more natural graphical user interface for toggling options and using menus is needed to complete the CAVE, something that I would like to explore tinkering with in the future.

My professor and head of the lab also ordered huge flat-screen three dimensional televisions to create a smaller and easily transportable CAVE system which is a great idea. The CAVE system is quite large and requires a lot of room and maintenance (calibration, projector issues, etc.) where as three televisions at a fixed angle would almost give the user a 180 degree field of vision!

Pick up the Pieces and Run

After learning the basics of OpenGL and the Ogre3D framework (as well as Visual Studios 2013), I was given the task of continuing the experiment of a previous researcher at the laboratory. Unfortunately, while the experiment was in a working condition before he left the code needed to be fixed due to version control issues. After a little direction from it’s original writer, I was able to set up a new environment and successfully compile the code. It was a close to completion but not working quite right.

As I am unable to explain the problems I encountered fully just yet, I will say that digging deep into the documentation of OpenVibe, Emotiv Epoch, Ogre3D, and Visual C++ was extremely helpful. Recreating the development environment on a different machine was difficult enough (one of the few times I had to do this and delve deep into all of the somewhat misleading ways to include a file or library in Visual Studios). We had everything working soon enough and I was soon able to begin running the experiment with participants.

 

Mazes with Sketchup and Unity3D

Utilizing Sketchup (Google/Trimble) and the Unity3d engine, I have created new mazes and simple controls to navigate them. The sheer speed that Sketchup gives me is unmatched in regards to other three dimensional modeling softwares that I have come across. After building a one level maze and another multi-level maze, I began importing them in to Unity 3D. Collision detection is automatic when importing a Collada file from Sketchup as long as the collision detection box is checked.
After importing models into Unity3D (mazes in this case) it is then a simple matter of allowing movement within these mazes. A simple first person camera controller is a perfect built in solution for me but needed some work. In order to prepare the program for simple left, right, forward, and backwards messages the camera had to lose some of it’s superfluous abilities. After rewriting the camera control code in Javascript, the camera now simply makes 90 degree turns a forward and backwards steps. This was done to allow ease of use when using a Brain Computer Interface as a means of input.

Make Your Own Emotiv Epoc Felt Pads

The Emotive Epoc I have been working with can sometimes be troublesome to maintain. I found out that the device had not been used in quite a while and needed some care before I could begin working with it. The headset uses a sensor that is only ready to record when wet (the solution we use is simple saline/contact lens solution). To keep the sensors moist it the Epoc uses felt pads above each of the sensors.
These pads do deteriorate rather quickly and must be replaced. We first attempted ordering replacements but found that the Emotiv website had them listed as sold out. Rather than let this halt our progress, I was able to fashion a solution using felt pads at the recommendation of Dr. Banic. The felt pads can be replaced with felt pads that are used to keep furniture from scratching hard-wood floors. After cutting them down to the correct size and letting them soften in saline solution I was able to fit them into the Epoc and begin reading brainwave data again. These felt pads are much sturdier than the previous ones and I highly recommend them as a solution to all who are facing similar felt pad problems.

Emotiv Epoc (developer)

Emotiv Epoc (developer) (Photo credit: drbakkerall who are facing a similar problem.

Emotiv Epoc & OpenViBE

Despite some worry I was able to get the Emotiv Epoc headset working. Prior to my arrival at the 3DIA lab, it had not been in use for several months. Upon inspection I found that there was quite a bit of corrosion buildup around the sensors. My first suggestion was purchasing a new sensor pack but the Emotiv website listed them as sold-out. This was problematic since waiting for a new one to arrive may have taken too much time out of my limited eleven weeks stay with the team here in Laramie, Wyoming.

I then turned my thoughts on to how to get the sensors reading again. After scouring the Emotiv Epoc forums and finding similar questions posed to the forum administrators (several of whom actually worked on developing the Epoc) I was able to find several methods of cleaning the salt buildup from saline usage (a requirement for the sensors to be able to read brainwaves). After several rounds of dismantling, cleaning, reconstructing, and testing, I was finally able to have all of the sensors reading again. Hopefully the headset will stay fixed, saving us time and money.

If you are attempting to clean buildup out of the Epoc please make sure you only clean the convex side. This is important because of the special polymer on the concave side that minimizes noise and allows for clear readings is very sensitive.