Group Actions
Engagement
  • Author Author: ciorga
  • Date Created: Date Created
  • Views 401 views
  • Likes 0 likes
  • Comments 4 comments
Related
Recommended

Smarter Life Challenge - Brainwaves based appliance controller - Update 3

ciorga

Here I am with my third update on my brainwaves based appliance controller project.  Things started to be harder and harder: after getting enough familiar with PSoC Creator and building a few small applications (that was fun!) I felt that I need to start looking into how I will measure brain waves. This is probably the most difficult part of my project, so I spent a lot of time reading and then I built my first set of sensors as shown in the figure below.

image

These are called passive sensors and are applied directly to the skin (usually using a conductive gel in between, but I try to avoid this if I can make them work without gel).

 

Next step I can describe as a tedious work of wearing the sensors and measuring brain wave signals in various parts of the head and process the data using various settings for signal filters trying to remove measurement artifacts especially the 60Hz component.  You can see me below in the middle of these experiments (and you can imagine me later when I got even more tired – not to mention how many times I asked my self the question why I did not pick something simpler for this project so I could be almost done by now…)

image

 

Measurements showed a lot of noise but I was able to sense when I closed and opened my eyes, as I show in the screenshot below.

image

 

Further analysis of brain waves did not distinguish any other type of brain waves, so I decided to improve my sensors by adding an amplifier.  The second set of sensors that are now called active sensors are shown in the picture below:

 

image

With these sensors I was able to measure a combination of alpha (~8-12Hz) signals on the occipital part of the head, as you can see in the FFT screenshot below.

image

image

 

The signals are still hard to extract out of noise so I don’t know how I will be able to distinguish among various types of brainwaves and then to use them to control appliances.  This is a hard challenge for which I do not have a solution yet.   So this is where I am now; basically the measurement of brain waves remains still a challenge and this is what I will be focussing on next.

 

All my best wishes until my next update, and good luck to all smart challenge contestants in your project work.

Cosmin

  • Hi Cosmin,

    Very excited to see this update! very space-age technology with all your brainwave sensing =)

     

    A couple thoughts:

    - you could consider having the PSoC 4 in the sensor itself, that way you could get rid of the amplifier you have in each sensor by using PSoC 4's Opamps. maybe helps with splitting the processing across 2 devices as well.

     

    - there'a a PSoC 5LP on the Pioneer kit, which is the big brother of the PSoC 4 chip. Its Cortex-M3 based for some additional processing power and additional analog capabilities. But more importantly, the PSoC 5LP has a DFB (digital filtering block) that is basically a math co-processor that will allow you to filter out some of the noise issues you're seeing. There's also a very nice Digital Filter Component in Creator that help design filters with a simple GUI. here's more on the DFB: http://www.cypress.com/?rid=59626

     

    good luck on your next steps!

  • Hi Cosmin,

    Most consumer 'brainwave' detectors work by in fact monitoring your facial muscles; using more EMG than EEG, although most do not admit that. If you're pressing hard enough to think really really hard, you'll start squinting, and taddaaaah.... Brainwave (not! it was the squinting....).

    The hard thing about EEG is that your skull and the fluid around your brain is actually shielding quite well. Although I appreciate your determination to not use gel, I would first skip that requirement and use some gel or moist to get good signals. Having done an EMG project in the last few weeks has shown me how much a bit of moist changes your signal. Using sandpaper to eliminate the upper skin layer improves the signal even more, but I understand when you don't want to do that :P

  • That's still pretty awesome. I wonder if it could be used to detect REM sleep. What do you get if you move your eyes while they are closed? Can you try data logging while asleep to see what you get?

  • Hi Cosmin,

     

    Have you looked into some of the commercial head sensors?

    I picked up a head mounted system for a couple of hundred USD that included software to process the basic low frequency levels for identifying changes.

    Another outfit has a head scarf type arrangement that also used dry sensors for collecting brain data.

    Unfortuantely, I cannot remember the names, but I will check to see if I can find the information.

     

    Good luck,

    DAB