Slashdot Mirror
NASA Controls Jet With Nerve Signals
__roo writes: "According to this press release, NASA scientists were able to control a 757 jumbo jet simulation using neurolectric machine control -- muscle-nerve signals fed to a computer, which used a neural net to learn how to interpret the signals. The first prototype armband was made from exercise tights, and used metallic dress-buttons as dry electrodes. This page has high resolution photos of the device."
I can't belive I'm the only Anne McCaffery fan here. Surely someone else must remember Helva, Simon, and Tia. Here we are getting closer to the day when cripples can scout around the universe and noone remembers Science Fiction has perdicted it.
um - little known fact:
(would YOU publicise this?)
Fighter pilots on long (8hr+) missions have been known to wear adult diapers.
These are my friends, See how they glisten. See this one shine, how he smiles in the light.
The neural net would allow one to almost directly map the human musculature to the airplane control surface...
Why did I suddenly picture the pilot running around, holding his arms straight out to the side "flying"?
By definition, all theory is not "real". Macross plus had some interesting theories.
I am a bit confused on the moderation, 3? This comment is not relevant to the story at all.
One of the elements of the show is that there is a design competition between two designs, the YF-19 and the YF-21. One of the things that the YF-21 design had going for it was that it had an interface similar to the one mentioned in the story. The main difference is that the YF-21's interface didn't require the pilot to actually move any part of his body. He could simply "visualize" the whole aircraft as an extension of his body, and control it that way.Another thing worth mentioning, there was a scene where the YF-21's pilot visualized an easy way to kill the YF-19's pilot given the current circumstance. The YF-21's computer system took this as an order, and they YF-19's pilot was almost killed. This illustrates a potential danger in these kind of systems that the designers and/or users will have to be weary of.
I wonder if people who haven't learned to integrate their activities with a machine at an early age will ever manage to have this happen as seemlessly as you suggest. The people who invent this stuff may end up using this stuff rather awkwardly, like someone speaking a late-learned language, while their children will truly be the ones that inhabit a new world.
For some reason, I find the whole concept of this rather unsettling. I guess I don't like the idea of becoming overly dependant on a machine. What if the power goes out?
-- It only takes 20 minutes for a liberal to become a conservative thanks to our new outpatient surgical procedure!
I am curious if anybody knows how hobbyist friendly this field is. It looks like the physical hardware needed to pick up these signals isn't much, but i can imagine you need some _REALLY SPIFFY_ (read expensive, large, current-hungry, heat dissipating...) amplifiers to make this sort of thing work, and then some really spiffy algorythms on the computer side to filter out noise that gets picked up, plus whatever low-level chatter is happening on that nerve...
So here is my question: Does anybody that knows something in this field know of a source of information on this? Is the technology patented by somebody? How complex are the electronics, and are the algorythms for extracting the data public?
The reason i ask is because i've always wanted to be able to do this, now i don't want to land jets, or play quake, i have a much more modest application in mind, i'd like to be able to get several (as few as four would still kick ass, although up to 10 would be nice...) reproducable (you don't have to be able to reproduce them from memory, there will be a visual feedback mechanism, so you know what you are inputting now, and you can watch it change as you "move"). I don't really have a practical use for this, and i'm sure i'll _always_ be able to type faster than i can use this method, but i still think it's very cool.
So i guess if it's something i could concievable do for under $1000 assumming i already had microcontroller tools, a scope, prototyping tools, etc... (so i'm talking only parts, books, software that must be purchased, and oddball tools...)
Thanks, and i really hope i hear from somebody, because this has sort of been a dream of mine since i was about 12 years old =:-)
---
Play Six Pack Man. I
Control a Jumbo jet with nerve impulses...
Stewardess bends over to pick somthing up, pilot gets a woodie, and plane goes into a nose dive!
:-(
Course.. the pilot could start playing with himself.... watch that plane go up and down, up and down.... :)
more recently I saw a product hyped up to "read your mind" you placed your finger on something, and was able to move a plane just by "thinking" about it. when in reality you we're subconsciously moving your finger the way you wanted to go, this kind trick is common in other applications.
The use of a neural net however is quite good. neural nets are currently used in speech recognition, and writing recognition. basically you say "here's some data, it means A", "here's some other data, it means B". the neural net will be able to tell the two apart and allow for a good degree of error. this is the jitz of it, I'm not really a student of the field.
so it makes sense to use a neural net for a task of "these muscles patterns mean move left", and so forth. I'm just surprised I didn't hear about that success of such an application till they landed a freakin jet with it! but then again slashdot is eregular about there coverage of things, i imagine cmdrtaco and gang turned down the previous articals leading up to this one.
-Jon
Streamripper
this is my sig.
- tokengeekgrrl
"The spirit of resistance to government is so valuable on certain occasions
As for actual piloting or other safety-critical applications, though, I have to admit skepticism. Anything where fidgeting could actually result in death should probably be discouraged.
OK,
- B
--
http://www.bradheintz.com/
- updated
I have a few concerns with that type of control: 1. What about muscle spasms or craps. Plus what if you need to "stretch" those muscles a minute? Do you now have to require co-pilots in every aircraft (course commercial aviation requires this anyhow)? 2. Hypothetically. If the pilots are incapicitated, how is someone else going to *easily* take control of the aircraft. Oooopss wait while I strap on some electrodes... oh nevermind there's the ground.
This new technology is significant in that neuroelectric control of computers can replace computer keyboards, mice and joysticks for some uses
It's been stated flippantly by some other posters, but I really seriously hope none of these uses include safety-critical applications, either for the user or for the user's "clients" of whatever sort. The example of an airline pilot is a good one--is it really smart to think that a pilot will be able to keep his hand under *complete* control for the duration of a 4 hour cross country flight? Much less the much longer intercontinental flights? Can you imagine the difficulty of keeping your arm completely still during the period of stable flight? I suppose it wouldn't be *that* hard to cut in and out with the autopilot, but still....
Furthermore, one advantage sticks etc. have is they don't require you to be physically tethered to the control system. If a pilot today has some medical emergency, not only does the copilot have his/her own stick, but the pilot could be removed and any other person capable of flying the machine could very quickly take over. How long would it take 1) to move the electrodes and 2) train the neural net for another person?
This really does not seem like it would be a good technology for any kind of control system where you have human failsafes to protect safety.
7 November 2006: The day Americans realized corruption and incompetence weren't addressing 11 September 2001
"I'll bet you $100 that I can, too land a plane with both hands behind my back!"
-*- Any technology indistinguishable from magic is insufficiently advanced -*-
Sheesh - the use of quotes in that article would make Dr. Evil proud:
...ability to control and "land" a simulated 757...
...control an aircraft without the aid of a "stick"...
...a simulated "damaged" aircraft...
Scientists outfitted a "pilot" with an armband...
Hmmm... a little over doing it? Do you need to put the word "pilot" in quotes (well, I did just then, but then I needed to, because... Oh, forget it.
OK first of all, there's a huge difference between what NASA does research on, and what the FAA approves. The FAA is a very conservative, safety-oriented organisation, and avoids change as much as possible. If this thing was put into fighter production tomorrow, it would be a decade before civil planes saw such a device.
Secondly, let's look at what this is: a fundamentally new way of controlling a plane with the same old movements. It's exciting and innovative, but effectively the pilot is still flying in the same manner as he did before, although without a joystick in his hand. Fly-by-wire systems and positional sensors offer the same capability.
Ultimately though, this is the thin edge of the wedge. Make no mistake--this will lead to entirely new ways of interfacing with machines of all types, and may be the start of true virtual reality. (like the transistor was the start of the modern portable computer) This isn't a device - it's a technology.
"People who do stupid things with hazardous materials often die." -- Jim Davidson on alt.folklore.urban
All the time, he wonders when NASA will learn to convert to and from the English and SI system.
You know, this whole idea sucked really bad the first time when it was called a POWER GLOVE.
...All I can say is that my life is pretty strange...
The Anime Geeks will note that one of the Valkyrie fighter jets/mecha in the Macross Plus series was controlled via electrodes and biofeedback.
The next Slashdot story will be ready soon, but subscribers can beat the rush and slashdot the links early!
The consequences of direct neural tapping are mind blowing, and in more than one sense. One question is that when everyone has their brains wired up to this tech and to the internet, everyone becomes educationally equal, and everyone can learn new skills very quickly. This technology is a step in the direction of a classless society.
You know exactly what to do-
Your kiss, your fingers on my thigh-
You know exactly what to do-
Your kiss, your fingers on my thigh-
I think of little else but you.
Can you imagine the pilot having too much coffee?
"Sorry folks for that wee bit of turbulence. I drank an extra cup of coffee today and I'm a little jittery"
Probably going to be the first plane crash due too caffiene.
Er, no. This is about muscular nerves. Picking out eight myoelectrical signals through (several?) cm of meat is no mean feat, but distinguishing what's going on in millions of ganglia through a skull? And that's merely considering the scale difference; I dimly recall the idea that muscular nerves were qualitatively different than the ones which we think with.
-*- Any technology indistinguishable from magic is insufficiently advanced -*-
It would be just like walking... at first, interacting with the computer pack would be awkward, but after a while you'd do it without thinking, and it would become a part of you.
Just as the avilablility of an always-on DSL connection allows people to use mapquest rather than storing an atlas at their house, this technology will allow humans to forget the millions of trivial facts and focus on understanding and mastering skills.
--