Slashdot Mirror

← Back to Stories (view on slashdot.org)

Controlling a Robot With the Emotiv EEG Headset

Posted by timothy on from the bad-way-to-steer-a-car dept.

robotsrule writes "This video shows a WowWee Rovio mobile spy robot being piloted remotely over a Skype video call using thought, facial gestures, and head movements, with the help of the Emotiv Systems EPOC 14-electrode EEG headset with built-in gyroscope. An accompanying article explains in detail how the system works and what steps were taken with the Emotiv systems SDK to interface with the EPOC headset along with code samples. The system is based on Robodance 5, a free robot control program for consumer robot owners that will be out in beta on May 20, 2010."

21 of 33 comments (clear)

  1. Innovation times... by vikingpower · · Score: 1

    ...are here again. Yee haw !

    --
    Religous speak to God. Insane are spoken to by God. When all shut up, one can finally hear Shostakovich in peace
  2. Blah blah blah by BadAnalogyGuy · · Score: 5, Funny

    *insert bland comment about applicability to military purposes here*
    *insert stupid comment about applicability to porn here*

    1. Re:Blah blah blah by Errol+backfiring · · Score: 2, Interesting

      I think any soldier prefers a trigger to operate rather than accident when he gets the hiccups. And there are already "augmented reality" systems in military use for quite a few years now. Pilots can aim with their helmet in some aircraft.
      I see more applications for people who can not use their arms, either because of safety (say, when operating hydraulic presses or other carnivorous machines), because they have their hands full with something else (like helicopter pilots), or because they are disabled.

      --
      Nae king! Nae laird! Nae yurrupiean pressedent! We willna be fooled again!
    2. Re:Blah blah blah by u17 · · Score: 1

      *insert bland comment about

      ah fuck it

    3. Re:Blah blah blah by DeadDecoy · · Score: 1

      pfft, I was wondering if it could get me a beer, without me ever having to leave my basement.

  3. Better than Mouse+Keyboard for FPS? by purpledinoz · · Score: 3, Interesting

    I wonder if this would give people an edge in Counter-Strike compared to the regular mouse+keyboard setup. If so, I hope those 12 year olds don't get a hold of it. CS:S is hard enough already!

    1. Re:Better than Mouse+Keyboard for FPS? by imakemusic · · Score: 4, Funny

      Whereas Counter-Strike is serious business?

      --
      Brain surgery - it's not rocket science!
    2. Re:Better than Mouse+Keyboard for FPS? by L4t3r4lu5 · · Score: 2, Interesting

      Sure it will, as long as they can maintain an expression of "cautiously aroused" while staring directly up and with one eye shut, jaw clenched.

      Be wary, though; If you shut the wrong eye, you drop all of your grenades at your feet and teabag them.

      --
      Finally had enough. Come see us over at https://soylentnews.org/
    3. Re:Better than Mouse+Keyboard for FPS? by wjousts · · Score: 2, Insightful

      We're not comparing Emotiv to Counter-Strike. We are comparing Emotiv to a mouse and a keyboard. Emotiv is a toy with some great marketing spin. A mouse and a keyboard are tools that you can use to get stuff done (including playing Counter-Strike).

    4. Re:Better than Mouse+Keyboard for FPS? by Jaqenn · · Score: 1

      My understanding is that there's a delay of about 200ms between your brain saying 'click the mouse' and your hand doing so. So there's our baseline.

      The brain-scanner has about a 1-second delay before it can 'settle' on a pattern. Useless for people who still have their hands, IMO, though their website claims it's good for tertiary commands like 'show the minimap'.

      The facial recognition is supposed to be very fast, and the delay between your brain saying 'twitch your eyebrow' and your face doing so is ~40ms. Good for competitive players, I guess, but I don't think it's worth $300.

      I do software development for a living, and in my idle quest to avoid RSI I sometimes wonder about replacing my mouse with something else so that I don't have to move a hand between the keyboard and mouse all the time. I'm curious to know if the gyroscopic sensors on the headset can be rigged into a functional mouse replacement (either for gaming or for dev work), but I've never seen any articles exploring it. I'd value any insight anyone else can give on this topic.

      --
      You are awash in a sea of fiercely stated opinions. Obvious exits are: 'File->Quit', 'Reply', and 'Page Down'.
  4. cool...one step closer by hesaigo999ca · · Score: 2, Interesting

    One step closer to getting exoskeletons built and controlled by these headsets to allow people with spinal injuries to walk again.

    1. Re:cool...one step closer by Anonymous Coward · · Score: 1, Funny

      One step closer to getting exoskeletons built and controlled by these headsets to allow people with spinal injuries to walk again.

      I was going for mind-controlled robot girlfriends who will do whatever you want but, yeah, I guess your thing would be pretty good too.

  5. I for one... by kiehlster · · Score: 1

    welcome our new mind-controlled robot overlords.

    1. Re:I for one... by Mindcontrolled · · Score: 1

      You called?

      --
      Ubi solitudinem faciunt, pacem appellant.
  6. hardly EEG by Anonymous Coward · · Score: 4, Informative

    Almost all of the degrees of freedom come from head motion and muscle artifact. EEG is very sensitive to facial muscle artifacts, and when you actually record EEG the patients have to keep very still. I've even read of EEG/TMS studies where they use botox to paralyze the scalp muscle. So every time he clenches his jaw, he obliterates the EEG signal over the frontal cortices (the part he is supposedly recording). Basically this whole apparatus is just a fancy, commercial electromyogram (EMG). You can build an EMG with off-the-shelf electronic pieces (i.e. resistors and capicitors) for probably around $10. There's essentially no EEG involved here, except making a headline. Using EMG to control a robot is done every semester in hundreds of undergraduate electronics classes.

    1. Re:hardly EEG by Cycon · · Score: 1

      Almost all of the degrees of freedom come from head motion and muscle artifact. EEG is very sensitive to facial muscle artifacts, and when you actually record EEG the patients have to keep very still.

      The larger problem with the Emotiv EPOC headset is that the EEG sensor locations it provides do not match up to where "real" Brain-Computer Interface (BCI) research is focused. So even if you wanted to do control by "pure thought" alone the best-known areas of the brain where these signals are located are not measurable by the Emotiv EPOC.

      Electrode placement is based on an international standard called the "10-20" system:

      http://www.bci2000.org/wiki/index.php/User_Tutorial:EEG_Measurement_Setup#The_10-20_International_System

      Most BCI applications focus on "imagined" movements around the right arm or hand, left arm or hand, and feet. The parts of the brain which produces electrical signals when neurons related to these extremities fire are located in the C3 and C4 sections of the top of the scalp in the diagram at that URL. Another important location is the "Cz" sensor at the exact top of the crown.

      Unfortunately however, the key Cz, C3, and C4 electrode locations (going by the 10-20 scale) right/left/feet motor control are not available on the Emotiv hardware. Instead their hardware provides electrodes in the following 10-20 locations:

      AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4

      My understanding, based on discussions with Emotiv, is that they designed their headset with as many unique channels of information as possible, at the best price/feature ratio, which would fit the most number of potential users in a one-size-fits-all form factor. This last restraint prevented them from Cz, C3, and C4 because the exact locations from user to user were not consistent enough to be relied upon in a consumer setting (their target market). Locations for an adult would not be the same as an adolescent user, and getting the locations lined up precisely is "too hard" for the casual or non-technical public.

      So in other words, if you want as fine-grained control by thought alone as the current state of technology allows, you'll have to wait for a updated EEG headset model from Emotiv or another manufacturer - or of course learn how to build your own from the .

      --
      Your Brain + EEG + LEGO Robots = Brainstorms
    2. Re:hardly EEG by Cycon · · Score: 1
      Interesting. I posted much of the above information to the original article, and the content of the post now appears to have been censored:

      cyconx

      April 27, 2010 01:17:59 GMT

      \

      --
      Your Brain + EEG + LEGO Robots = Brainstorms
  7. Re:Worried Sad? by dsavi · · Score: 2, Funny

    In the future: Everyone will steer robots with facial gestures and concentration. And be depressed.

  8. t2 by windcask · · Score: 1

    Miles Dyson from Skylab was not available for comment.

  9. Similar Open Source project for LEGO Mindstorms by Cycon · · Score: 1

    Here is a similar project using the same Emotiv EPOC headset to control LEGO Mindstorms robots via EEG:

    http://brainstorms.puzzlebox.info/

    YouTube video link

    (disclaimer: yes, I'm the project lead)

    --
    Your Brain + EEG + LEGO Robots = Brainstorms
  10. have been working with one of the EEGs... by vuffi_raa · · Score: 1

    I have been working with the neurosky headset as of late, it is less accurate (not in readings but since it has less contact points id has a more broad measure of activity) and doesn't measure facial features but is as easy as putting on a set of headphones and way less expensive. The emoviv set seems like it could be really cool in more of a controlled setting but I don't see it hooking up to your home console with much success anytime soon.