Slashdot Mirror
NASA Snake-Bots
faqBastard writes: "NASA's been working on some pretty cool snake-bots for exploring outer space. All kinds of neat features and capabilities ... " Robotic snakes certainly seem to be slithering into our future. OK, they look practical and intriguing -- but they give me the willies.
Reminds me of the constructor at soda.
It is only a slight exageration to say the science by definition is unrealistic, and engineering is realistic.
If the Wright brothers had stuck to realistic science, they would never have flown.
If Einstein had stuck to realistic science, he would have been a patent examiner for the rest of his life.
If Newton had stuck to realistic science -- now that's an interesting thought; would that have meant alchemy back then?
Yes, all this is a simplistic exageration. But complaining about unrealistic science is simplistic exageration itself.
--
Infuriate left and right
In the meantime, NASA will contact the NSA about hundreds of prank calls to Mission Control - apparently hundreds of people are calling in and breathing heavily or asking for phone sex. Oh, and like many other slashdot articles, it'll be filed in triplicate on the front page, and then the other copy (or two) will spontaniously disappear without a word said..
(note for the humor-impaired: the above post was intended to be funny, not a troll).
I haven't been around there since winter break, so I don't know the current state of things exactly... that was one of the issues being looked at during my winter break, and surely something that would need to be improved vastly for space exploration.
a tether has advantages, however -- especially when it tends to go wandering down hallways on its own (I was programming an older version, with a HC6811 as the brain, and exactly ZERO sensors) and when you have it climbing ladders and such.
Lea
heh. I've pointed him at /. enough... if nothing's coming up, he might, though of course I couldn't speak for him. wheedle a bit, though, yes :)
Lea
I think you're vastly underestimating the amount of current it presently draws... however, if you're in free-fall, you wouldn't need such hefty motors, which would solve most of that problem. the problem with that solution is that from what I've talked with the NASA guys about, they were thinking also surface applications (they make some pretty sweet freefall robots too, but mostly mars rovers). the one that NASA's got (unless they rebuilt it since I've seen it) is made out of PVC and little servos -- it's quite weak. the newer polybot (there's pictures of it on the parc page) is VERY heavy, with hefty motors to move it around.
The snake will need to perform complex folding and grasping motions using a dozen joints. Designing that control system will be formidable
you're telling me! it's really very interesting controls research stuff. what do you do with a few hundred little modules which are pretty useless by themselves, and autonomous? how do you get them to work together? what happens when some of em break? (because some of them will break, if you have that many) if it wasn't so hard, the researchers probably wouldn't be so interested (I know I wouldn't) :)
And then they want detachable segments!
indeed. the new polybot can do it, and in fact demoed it at Comdex.
the technology is facinating, as are the controls issues that accompany it. it's working incredibly well, I think.
Lea
What makes these snakes better or worse than Mars Lander for smashing into nearby planets with?
1. they can curl up into small spaces and make themself more resistant to high-G environments
2. they can change configuration depending on terrain
3. more robot per space/load becasue they can turn into someting much larger
4. resistant to breaking down, since they're massively redundant
for other reasons, take a look at my other post for the link to PARC's site.
Lea
well, I'd have to say that for something made "quickly and easily" there was quite a lot of effort expended. it's a quite impressive piece of hardware, though if you look at the goals on the webpage you can see that no one thinks that this is going to be the last generation, with motors with giant gearboxes and such.
/three/ generations in the space of 15 minutes (not keeping in mind that the easy availibility of compilers makes us a bit sloppier, since we know the compiler will catch mispellings, etc fast). hardware takes so much more time (even the version nasa's got takes time -- trust me, I glued a LOT of those kind together last summer) before you can even start debugging.
there is a certain difference between coding and "real" engineering in the time for feedback. I write code. I compile. I fix bugs. I compile. I fix more bugs. I compile. that's
also something that I think the article glossed over is that the basic architecture of these robots is reconfigurable. take a look at the polybot on the PARC page -- polybot is shown as a spider, a snake, and a rolling track, among other configurations. not only can it make these configurations, but it can change between them by itself, and (for example) standing up as a spider from completely flat on the floor takes a lot more power than a snake does.
in any case, since you don't have your email up, feel free to continue this conversation by mailing me at chialea at nanorobotics.org
Lea
I'd really like to know...
Boffoonery - downloadable Comedy Benefit for Bletchley Park
This sounds like a more advanced version of the Sky Worker project at the Field Robotics Center at Carnegie Mellon University. Although this project was origionally supposed to be for a robot to maintain orbiting solar cells, it seems to have involved to working on a robot to build, inspect, and maintain space facilities.
like so many futuristic visions, this cool concept would, in almost any conceivable implementation, be too costly to replace the standard methods of cabling.
a thousand tiny robots = a thousand points of failure. even if these robots were able to cost $20 a peice, they would still be far far more expensive than a standard fiber optic drop and the labor to install it.
In Capitalist America, bank robs you!
a thousand tiny robots = a thousand points of failure.
Not really. In fact, if they're interchangeable, a thousand tiny robots = 999 replacements if one fails.
> The tiny snake, just 1 inch (2.5 centimeters) in
> diameter, could be used to inspect gas lines
> here on Earth as well. Marzwell said the snake
> could use the pressure of the gas within the
> pipeline to turn a tiny turbine to produce its
> own electricity.
Does anyone else here see the problem with sending an electrical generator into a gas line?
--LordEq
>Robotic snakes are easier to maintain than space monkeys!
Hey now, I resent that!
Or we could use robotic ants like in an earlier story: http://slashdot.org/articles/99/10/11/0659214.shtm l
:)
They can pull cables like real ants!
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
Hmm that could be a problem [grin]. :)
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
cables but if they can made thin enough they could become cables themselves. You could have self assembling structures. And if they are modular they could be able to join with each other to become longer chains or break up into smaller sections to search or repair in parallel. Cool...the possibilities are endless.
FarHat
At the intersection of computation and biology.
I attended a talk by Mark Yim at Stanford around last November. As I recall, there were still many problems with his transforming robots (nifty as the idea was). The biggest problem seemed to be power issues: in the video clip we saw, the robots had tethered power supplies - the motors in the robots were too energy draining for (reasonably-sized) batteries to be used. Have these problems been solved yet?
Since the big problem with looking for life on Europa is burrowing through 100km of ice, burrowing snakebots would be pretty cool, and they wouldn't require massive machinery.
Problem: how to dig quickly enough to get through that much ice without running out of power.
"You can't get something for nothing." - my grandfather, on the stock market and Reaganomics.
I don't see power as a problem in space: since the snake weighs nothing, it only has to act against its measly momentum (until the astronauts hold a snake race ;-). A simple latching mechanism, maybe with a spring to continuously exert force, would hold joints stiff for grasping. Solar cells, along with small batteries, would probably be sufficient. And if they're working in darkness, they can just use temporary mirrors.
But I do see control as a problem. True robotics, especially in space, has a miserable track record. Even apparently simple things are actually very difficult, as Deep Space 1 pointing its camera in the wrong direction demonstrate. The snake will need to perform complex folding and grasping motions using a dozen joints. Designing that control system will be formidable. (The space shuttles' vaunted "robotic" arm is no more robotic than a bulldozer. It is a waldo (remote-controlled hand) driven by a human operator at a joystick.)
If the snake is anything like the prototypes they showed, it will have lots of bearings. Vacuum welding will be a minor challenge, and thermal expansion could be a major concern. Many missions, such as Galileo, have been impaired or lost from solar arrays and antennas that were supposed to fold out, but instead jammed.
And then they want detachable segments! That probably means electrical connectors (unless they put a radio transciever in each segment, which makes it a networking nightmare). On every design project I've ever been on, connectors have been the single largest pain in the ass (picking a microcontroller or transistor is easy compared to picking a connector). And not only is it a connector, it has to attach/detach (in vacuum) under the supervision of a robotic (read dumb) brain.
I don't mean to put the snake projects down -- they just have so many compounded difficulties.
There is one way the snake beats everything else hands down: redundancy, both of operations and repair. Most spacecraft can't keep several full sets of spare parts in a bucket! When you're 100 million miles from home, that might overshadow the other shortcomings.
-- ;-)
Kuro5hin.org: where the good times never end.
I am waiting for that enchanting moment when artifficial snakes will be used as another form of recreational activity: -Oh, Al! -Your snake! It's slithering; now its sidewinding! -Peg, it was the best $150 I have ever spent, now I can actually get some sleep.
You can't handle the truth.
I wonder if they could be used to pull cable?
...................
...................
...................
Ever heard of the Mars balloon?
It was a big helium balloon which was supposed to move around Mars during daytime (with the help of winds) and during the nighttime it would go lower and stop.
The snake-relation here is the device attached to the balloon. It was a snake look-a-like, formed of interconnected metal cones containing electronics and measurement gadgets inside. When they tested the balloon, they set it free in France and it ended up in the USA, to be picked up by a farmer when the balloon ended stuck on a tree.
The snake was designed to be such that it wouldn't get stuck between rocks. In fact, the snake performed exceptionally well, slithering between rocks.
What happened to the Mars balloon and the snake?
Well, it didn't make it (the balloons were removed from the project), due to many reasons, one of them being the breakdown of Soviet Union and one being the failures of Mars probes. For instance, the Mars Observer carried the "Mars Balloon Relay" (MBR) and as we know, the MO disappeared 3 days before arriving on Mars. So no relays were deployed.
More information about the balloon, and here about "aerobots".
I couldn't find a link about the snake attached to the balloon, sorry.
however, about making it cheaper, a lot of what is expensive is custom (and very complex) circuitboards, processors that are very hard to find, and stuff like that. very sweet hardware -- but it's not going to be cost effective right now, or at least until it's a little more developed.
then again, Mark hasn't told me what I'm doing this summer (and I've been too occupied with finals to ask)
Lea
If you read through to the end of the article, they mentioned that the eventual hope is to replace the current system with something that uses a sort of electrical muscle substance. Some sort of thin, metal coated plastic that would deform in response to small electrical currents being placed through it, in much the same way as animal muscle works. Additional bonus over something that responds to PH is that your mechanical snake doesn't have a seizure if it wanders into some sort of natural acid/base deposit/
do robotic snakes eat computer mice?
-- Dr. Eldarion --
It's not what it is, it's something else.
No, this was not on Slashdot a year ago. Did you read the article on space.com?
There was a previous article on snakebots but this article is new and goes into a lot more detail on proposed uses for the snakebots and the benefits of them.
Think of it as an update, continuuity is good, slashdot should followup on interesting articles, like remember when we mentioned cool widget foobar? Well this is where it is now.
This general idea has been around for years; Gavin Miller has been doing snake robots and snake animations since the 1980s. (Miller's a great guy, but he has this thing for snake locomotion.) Snakelike robot tentacles have been built and used, with modest success, as spray-painting robots.
There's probably a cool toy in this. The technology needs to be redesigned by somebody like the guy who did the Furby to get the cost down, though.
Don't these people ever match movies? You create some sort of icky technological horrors like robotic snakes and deploy them in an isolated location like a space station or martian colony and they will certainly go berserk, killing all but the most charismatic male and his love interest. These two characters of course defeat the evil technology just in time to catch the last spaceship back to earth.
there are also a lot of related projects, such as Proteo and Digital Clay that are also very interesting stuff.
disclaimer: I currently work on this project at PARC (well, when I'm not in school), and I used to work for that group at NASA (for a summer).
Lea