Robotic Space Workers of the Future

Roland Piquepaille writes "In an article named "Puckish robots pull together," Nature describes the work done at the Polymorphic Robotics Laboratory (PRL) of the University of Southern California on self-reconfigurable teams of robots. There, Wei-Min Shen and his colleagues simulate the absence of gravity by creating a 2D representation of space by using an 'air-hockey table.' With jets of air flow blowing on the surface, the 30 cm-wide robots, working in pairs, evolve in a frictionless environment, pick elements such as girders to assemble structures like if they were in space. NASA will use these teams of autonomous robots to build space systems like 10 km-long arrays of solar panels and other huge spatial structures. You'll find more details, illustrations and references in this overview."

Comments

Is it truly frictionless?

[NightWulf]

It's an air hockey table! It's not exactly frictionless, as there is air resistence and other factors. Maybe these scientists know more than I do, but I can't really imagine an air hockey table can even remotely simulate space; where you bump into something and when you bounce back you'll keep going forever, etc. Other than that, it looks intresting of them all working together, a beowulf cluster of space robots, heh.

Re:Is it truly frictionless?

[quisph]

No, of course it's not perfectly frictionless, but it's "frictionless enough," you might say, to test the concept. Anyone who's taken more than a little bit of physics has probably done an experiment using "frictionless" air pucks at some point in time; this is nothing unusual.

A much bigger shortcoming is that this is 2-d instead of 3-d. But then, a ride in the Vomit Comet doesn't come cheap.

Re:Is it truly frictionless?

[kfg]

Maybe these scientists know more than I do. . .

I know that when we bought the air hockey table for the physics department we knew what we were doing.

You are correct, it isn't "frictionless," but it is a much closer approximation to frictionless than is, say, a shuffleboard table, which itself is fairly low friction as these things go. It is frictionless enough that if you were to build an airhocky table a mile long you could drive the puck from one end to the other ( for that matter a golf ball has been driven a mile across a frozen lake, which has both more surface friction and air resistence than a puck on an airhockey table).

Having so little friction that miles are inside the bounds of relevant behavior makes yards even more so and remember that Gallileo was able to deduce frictionless behavior by rolling crude wooden balls down crude wooden ramps. You can do this thing called "extrapolating."

Nor is space itself frictionless. It is close enough that one may discuss it in those terms when discussing certain phenomenom, but this too is dealing only in pragmatic approximations.

Stuff doesn't "keep going forever." Space is not empty. Energy is lost throught various "winds" and collisions, just like on an airhockey table. In the real universe "when you bump into something" you often lose energy because real collisions are not ideal, and even light loses energy when it "bumps into something" (like, oh, say, something vaguely blackish). The total energy of the universe is conserved (the universe itself being "the system"), but the total energy of individual objects is not.

The airhockey table itself is an example of this, the puck slows down because it bumps into things and transfers some its energy to that thing. Like the table itself. Which loses energy to the universe.

Think about it, and perhaps you will come to a smaller gap between what you know and what the scientists know.

KFG

Re:Is it truly frictionless?

[WindBourne]

Keep in mind, this was not about physics, but about controlling devices. This was a test of the AI and logic that exists rather than a validation of the mathmatics of how to manuever.

Re:Is it truly frictionless?

[SEWilco]

A much bigger shortcoming is that this is 2-d instead of 3-d.

"the girder is complete. what is this "spiral" word in your query?"

For the last time

[Anonymous Coward]

For the last time, meatbag^H^H^H^H^H^Hpeople...

Please, do not use the "R-word".

We prefer to be called Electronic-Americans.

Thank you.

end transmission.

working in space?

[xlyz]

it seems that in the U.S. people really dislike working. they are sending jobs everywhere.

Robots tested on Air Hockey table?

[IllogicalStudent]

Hmm, I just hope the robots don't stop working after they let 7 goals get by them.

Little buddy.

[GhostChe]

These hocky puck things remind me of another robot devoloped by nasa. They both like floating around in space, but nasa's has more of the little buddy going for it.

Potential issue

[The_Mystic_For_Real]

It says that these things are able to continue to learn and adapt. I am not an AI expert, but how many mistakes does it have to make before it figures everything out? I have yet to see a machine programmed with every facet of the instincts that might prevent disasters from unforseen situations. Of course, humans make their own mistakes.

Re:Potential issue

[Mostly a lurker]

I have yet to see a machine programmed with every facet of the instincts that might prevent disasters from unforseen situations. Of course, humans make their own mistakes.

Indeed, humans make their own mistakes. Darwin had a theory (mostly accepted) about how humans have develop the "instincts" that help them improve their performance and in some cases keep them alive. Some experiments in AI based on similar principles seem to have held promise (e.g. see When Robots Play Games). Perhaps the key is to have multiple teams of robots with slightly different designs such that an error by one team is less likely to be replicated by all.

Why send jobs to robots?

[Adolph_Hitler]

This is the worst thing we could do as humans. IF you want society to fall apart simply make the majority of humans useless. I mean if Robots can do what the average human could do, what the hell is the average human useful for? I guess its time to start slaughtering and killing about 6 billion useless people so we have space for these robots. Don't you agree?

Re:Why send jobs to robots?

[Fweeky]

Well, it's going to happen sooner or later; one day it might even be cost effective in the general case. I hope we can come up with a better way of dealing with it than you when it finally is; that's probably giving humans a bit too much credit though.

"what the hell is the average human useful for"? Who said we had to be useful anyway? We have to survive (well, not really, but let's pretend); whether we do that working our asses off or having fun while our technology does our work for us doesn't make a whole lot of difference.

Re:Robots should be outlawed MOD PARENT TROLL

[Nom du Keyboard]

I think automation should be outlawed.

Smells like a troll to me.

Ludites reading /. now. What is this world coming to?

Capitalism and Robots do not mix.

[Adolph_Hitler]

First I never said capitalism is good. What I'm saying is capitalism and robots can not co-exist. Humans become absolutely useless once robots become efficient. Yes at first robots increase jobs and productivity, but soon the knowledge and intelligence level required to continue to program/repair/ or stay above the robots will become too much for the average human to handle.

Can we all have A PHD from MIT/Harvard/Yale/etc? Competition with humans in the third world is enough, and the population keeps increasing every year meaning competition keeps increasing. How the hell are we supposed to compete with each other as 6 billion humans along with the machines?

I can see it now...

[robertchin]

"I am bender. Please insert girder."

I'm one of the researchers

[wildmage]

Hi folks. I'm very excited that this project is finally getting some attention. The concept is simple, but it has been overlooked for quite some while.

Let me give some straight facts through all this futuristic market speak in the articles and from my professor. Where are we now?

1. We are trying to do a proof-of-concept that a team of robots can indeed assemble structures together in a near-frictionless environment.

2. We are currently trying to build a triangle out of 3 reconfigurable beams assembled by a pair of tethered robots. With a triangle we can realize more rigid and useful structures such as trusses.

3. We are halfway there. We have achieved two-beam assembly with reconfigurable connectors and everything.

We have been working on this thing for almost a year, and one of the things you might be asking is why is this so difficult?

1. Main issue is connectors. You want to have connectors that can be automatically assembled together yet provided tight tolerances and carry heavy loads. These are often conflicting requirements and this has required a lot of tinkering to accomplish.

2. Reconfigurable connectors. These are connectors that not only automatically connect, but also automatically disconnect. Give the above requirements in 1 and this becomes doubly more difficult.

3. Precision control in a "near-frictionless" yet noisy environment. This is very difficult. Our positioning is kind of crude, our propulsion is non-linear, and the noise in the air-table is not predictable. We've been able to accomplish a lot of our results by using the tether to pull the two robots together and assemble the beams together with a rolling motion.

For those of you who are interested in seeing our latest results I recommend going to the media page at our lab here

The last video (which is surprisingly not up yet) is here

For future reference, the research involved in "evolving and adapting" has not yet been done. That is future work.

Thanks,
Jacob Everist
everist@usc.edu

Re:what I don't get is this

[Tailhook]

Global warming - I don't care whether...

Were talking about experimental robots that float around on an air-hockey table. Somehow, this guy manages to end up ranting about "global warming."

Isn't it about time for you enviro-spaz activist types to take your hysteria elsewhere? Go find a political or environmental site and rant there.

Thanks.