Apr 01 2009

Controlling Robots with the Mind

Honda announced recently that they have built a device that reads the electrical impulses and blood flow of a person’s brainenabling them to control a robot with their thoughts. Sounds cool – but actually I did not find this to be a significant breakthrough. Other research centers have already achieved similar goals – remember the monkey that learned to control a robot with his mind?

Honda claims that their innovation is in developing a device that is painless because is uses only surface sensory and not implanted wires – but New York researchers have already accomplished that goal too, several years ago.

So really, this is a non-story (Shhhh….don’t tell the press).

Honda is a leading developer of robotic technology. So I guess they command some attention whenever they show a robot doing something. It is also likely that because of their robotics expertise they will play a role in mind-controlled robots in the future. But we need to put this unwieldy and clumsy device into perspective.

Their “mind reader” can discern four states of the brain – when the subject thinks about raising their right arm, raising their left arm, running and eating.  This takes several hours of training with the subject. Honda showed a video of their Asimo robot lifing its right arm several seconds after a subject commanded it to with their thoughts. Honda says that their system is still too crude to give a live demonstration. Again – I’m not impressed. Not even ready for a live demonstration?

So they have crudely duplicated technology already developed years ago by other research centers. Sounds like they are fishing for investment money or PR. Eh – maybe I’m being too cynical.

News articles speculate about using this technology to one day drive a car with mental commands. There is nothing wrong with this speculation – but this is certainly not going to be an early application.

One thing to consider is that there is a certain neurological buffer between our intentions and our actions.  We have to physically move our arms to move the steering wheel. There is a threshold of intention that needs to be overcome. By operating the steering wheel directly with the mind (depending on the sophistication of the reading device and software) this buffer may be lost. Operation of the car may be at the whim of every stray thought or distraction.

For this reason operating equipment like a car will probably be a late application, only after the the technology is fairly mature. In fact, there may never be an advantage to bypassing the arms and controlling a vehicle with the mind. Unless there is a physical impairment, the arms work quite well. I file this under “cooking with a microwave” – just because we can do something in a new fancy way, does not make it better.

The real applications (like heating with a microwave) will likely not be realized until after experimentation with real-world use. Right now we can only speculate.

But having said that, my speculation is that this technology will be most useful for, and will be early adopted by, the physically impaired. If one cannot move their arms then such a device would be extremely useful. Right now crude devices can already be used to move a cursor around a computer screen.  Once you can operate a computer with your mind in sufficient detail – a world of communication and control opens up even to those completely paralyzed. Such applications are also more likely to attract research money. And – such devices to not have to be as portable for someone already confined to a large wheelchair or to a bed.

The real point of this news item is for Honda to announce to the world that they are now in this game. Sure – they overstated where they are in relation to other researchers, and were probably jumping the gun, having to rely upon a video rather than risk a live demonstration. But it seems they are poised to invest heavily in this technology, which means they may quickly become a real player.

Share

21 responses so far

21 Responses to “Controlling Robots with the Mind”

  1. superdaveon 01 Apr 2009 at 9:59 am

    I’m in rare disagreement with you. Even though you are right about the tech being relatively old in research terms, it’s still pretty new to the public as this sort of technology has not filtered down to any consumer level products yet. Therefore I think that Honda’s achievement is worth mentioning as it represents a step in the commercialization of a technology. This is an important part of research even if it has more to do with marketing than science.

  2. Karl Withakayon 01 Apr 2009 at 11:00 am

    I’ll be more impressed when Honda starts building robots that can find Sarah Conner.

    Honda has a more effective PR machine than the other researchers, and the media (and media consumers) loves anything with robots, especially bipedal robots. Anyone who follows science and technology news heard about this technology a while ago, and wondered what the big deal was with Honda’s announcement, “We’re doing it too, but with cool robots!”

  3. HHCon 01 Apr 2009 at 11:13 am

    Honda’s Asimo robot commanded by a human sounds like they intend to build the perfect chauffeur. 2009 Mercedes Benz and Volvos already have computerized braking. Lexus LS 460 and Lincoln MKS automatically parallel park!

  4. eiskrystalon 01 Apr 2009 at 11:16 am

    4 states, several hours of training and several seconds before it moves.

    Sorry but that really sucks.

  5. Watcheron 01 Apr 2009 at 12:12 pm

    Sorry but that really sucks.

    For one of the first forays into this type of endeavor, it’s probably on par with everything else. How did our first attempts at genetic modification go? Or the first heart transplant? New science is always crude before revision. Cut ‘em some slack :D

  6. DevilsAdvocateon 01 Apr 2009 at 2:13 pm

    With all these recent posts on brain chips, robots, and microdynamos, one wonders what Dr. N has going on in his basement lab…..

  7. artfulDon 01 Apr 2009 at 3:17 pm

    He’s working on a robot that serves its own purposes.

  8. Karl Withakayon 01 Apr 2009 at 3:29 pm

    All of this has happened before, and all of this will happen again……

  9. DevilsAdvocateon 01 Apr 2009 at 6:05 pm

    How much do we *really* know about this guy?

  10. terrenceon 01 Apr 2009 at 9:15 pm

    @superdave “…this sort of technology has not filtered down to any consumer level products yet.”

    Not true. I think that http://emotiv.com/ uses the same methods (and some other modalities as well). It was all over the tech and gaming news sites a few months ago, so this sort of thing already has some public traction.

  11. superdaveon 02 Apr 2009 at 12:08 am

    @terrence, that is a fair point but Honda is a much more well known company. In the eyes of the public, Honda doing something is a bigger story than emotiv (not that their product isn’t cool). Yes, honda might suck at brain machine interfaces but I think the bigger story is that a company like honda is even attempting to do it.

  12. terrenceon 02 Apr 2009 at 12:52 am

    @superdave: You’re certainly correct that the real story here is that it’s Honda doing the work. I didn’t mean to imply that emotiv’s gear is in any way better – it’s probably not. For what it’s worth, I’m personally not at all excited by this technology, I was just pointing out that products do exist.

  13. Phospholipidon 02 Apr 2009 at 7:13 am

    My money is on the first mind controlled robot being either a DARPA hand or one it’s european counterparts, like the SMART-hand. That’s based solely on the number of groups and dollars pumped into the field, not the likelyhood of anyone actually wanting a prosthetic robot hand.

    There’s also the Donoghue group and their experiments with Utah-type arrays in the motor cortex of paraplegic patients that can (sort of) control a mouse cursor. The quality of the research can be questioned, but similar to Honda getting attention because they’re big, doing mediocre science on humans instead of monkeys or cats gets you published.

    I saw a slide presentation a while back of an implantable chip that could take hippocampal input and mimic CA3 neurons to the extent that mice could have their memory turned on or off. That’s probably the coolest mind-interacting device I’ve seen yet.

  14. Watcheron 02 Apr 2009 at 10:47 am

    All of this has happened before, and all of this will happen again……

    Can I be the one who gets to be haunted by the cute angel?

  15. HHCon 02 Apr 2009 at 11:35 am

    For Watcher, a cute angel smiley. o:-)

  16. Sastraon 02 Apr 2009 at 6:37 pm

    A while back, my father — who is a firm believer in the paranormal — ran across an article which talked about a device which allowed a person to control a computer cursor with their mind. He was gleeful when he brought it up to me, seeing it as scientific vindication for the existence of ESP. I had a hard time convincing him that, if anything, such devices support mind-brain materialism, since it all has to do with mechanical connections to neurons.

    I think he still believes this is confirming ESP.

  17. HHCon 02 Apr 2009 at 7:41 pm

    Sastra, Daughter’s mind-brain materialism = Father’s ESP. So there is nothing extra sensory in his definiton.

  18. ccromeon 07 Apr 2009 at 7:53 am

    I notice in this blog entry, and in the SGU podcast, you didn’t mention the proprioceptive system at all. I remember reading one of Oliver Sachs’ books, and he tells the story of the lady who completely lost her proprioceptive system, and *only* her proprioceptive system, due to an allergic reaction to anesthetic.

    After loosing that system, she could only make her muscles perform full on or full off. There was absolutely no subtlness to her movement. Every movement was with 100% muscle flex.

    Eventually the lady learned to use her eyes as part of the feedback loop to replace her proprioceptive system, and she could, albeit jerkily and clumsily, walk and sort of function again.

    It seems that using a ‘mind-reading’ system, you have the feedback path that is completely missing. It’s going to be hard to find a replacement path for your proprioceptors. Eyes are already busy. Ears are not so good for the purpose (you could use an auditory feedback loop I guess). You get the idea. None of the ‘regular’ senses will serve as very good proprioceptors.

  19. astronomyscience.neton 27 Apr 2009 at 11:32 am

    astronomyscience.net…

    I agree that if this becomes just the old model of CRM then the value of these types of interactions will be minimized. However, I think the point is more to build a community which includes the full spectrum of user types. When you can get your friend…

  20. [...] general, I’m in agreement with most of the points made in this NeuroLogica Blog post by Dr. Steven [...]

  21. [...] April of this year I wrote about Honda’s press release announcing that they had made some breakthroughs in the area of mental control of [...]

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.