Building robots on the moon.. how close?

[kelvin_zero]

Hi, Are there any estimates on how close we are to being able to build a robot factory on the moon that is advanced enough to build more robots from almost entirely moon resources?<br /><br />What interests me most is the possibility of exponential growth of industrial ability at zero additional cost after the initial investment.<br /><br />Simplistically, If a robot factory could duplicate all it's components in a year then it could increase itself a thousandfold in a decade, a millionfold in two decades, and soon the moon would be one big solar powered construction yard. You could upload new designs to the moon over the internet. You could build massive habitats before you bothered to send people. Building a magnetic launcher would be trivial with all the moons resources. We could start building all those 1960s style space habitats on whim. Exponential growth is just cool and even scary.<br /><br />What is the current state of play of this sort of technology? any estimates on how close we are or how big the first factory would have to be, or how long it would take to double in scale?

 

[crossovermaniac]

Not for a long time. Just think of all of the components that go into a modern car and the material needed to assembled them. Steel for the frame and cover (iron, coal, and limestone), copper for the electrical system, lead for the battery, acid for the battery, rubber for the tires and insulation for the wiring and plastics for the miscellaneous parts (petrolum-based). And all of this has to be gather together, refined and processed, and assembled into a car. Now try that with a robot. We're decades, at least, from self-assembling robots. However, automated robots that build one kind of product, let's say, photovoltaic cells, aren't that far off and has been proposed to build a mass solar power plant on the moon by David Criswell, though, there will still need to be a manned presence on the Moon.

 

[kelvin_zero]

I have to agree, at least on an empirical level, or else we would have done it by now.. but here are some niggly counterpoints:<li> Im not talking about self assembling robots, but factories, containing many different automated robots building just one product (solar cells, aluminium sheets) and also robots for mining and assembling.<li> In a way we almost already have these factories on earth. We mine with machines and assemble with robots, and many stages of smelting are far to dangerous for humans to get near. If someone can control a digger from the cocpit, they could control it by remote also.<li> So 'not for a long time' could almost be restated as 'not with a large industry of xxx size'.. (its not that simple, since we cannot use all the same techniques there naturally)<li> One more point is that robots on the moon do not need to compete with earth equivalents for efficiency.<br /><br />That last point deserves its own set of bullets!<li>A wheel of aluminium or even glass might work on the moon, if you have no rubber.<li>Building batteries sounds hard, but what about primitive capacitors, ie thin sheets of metal<br />holding static charge?.. crap, maybe.. our first robots might only work five minutes then rest an hour.<li>How much plastic do we really need? We use it very freely on earth for example to insulate wires. We can insulate with glass and sometimes just by making sure the bare wires do not touch.<br /><br />Finally<li>Some elements can initially be brought from earth like vitamins, eg computer chips. If these vitals can be reduced to only 0.1% of the total mass, then a 1 ton investment can still grow to a thousand tons before needing more input. With my simplistic rapid-growth agument above, it would still take 10 years for this thousandfold increase.<br /><br />FINALLY finally,<br />You have a good point about the solar powerplant. The first stages of building this robot factory might be to concentrate on getting power for a few years, nothing else</li></li></li></li></li></li></li></li>

 

[thereiwas]

In one of his interviews Mr Bigelow pointed out that the moon is a very hostile environment for moving mechanical parts. The "dust" is more like fine glass shards and it gets into everything. His experience in construction, where those big yellow machines need almost daily maintenance after exposure to just earth dirt, makes him wary of doing anything like that on the moon. Hence his new burying technique, and assembling the lunar station in LLO. <br /><br />The Mars rovers are having problems with the dust there as well.<br /><br />Bigelow's point is you can't just ship Earth-based manufatcuring technology to the moon and expect it to work for long.

 

[saurc]

Yes, you can't expect Earth built machinery to operate on the moon, if it is NOT DESIGNED TO. The Mars rovers etc, experience trouble, this is because the designers did not know how much dust would be there etc., but now that we have a lot of data on the nature of the environment on the Moon and Mars, we can hopefully expect machines can be developed which counter the problem.<br /><br />And anyway, if you can't bring machinery from Earth, where would you bring it from ???

 

[saurc]

>Simplistically, If a robot factory could duplicate all it's components in a year then it could increase itself a thousandfold in a decade, a millionfold in two decades, and soon the moon would be one big solar powered construction yard. You could upload new designs to the moon over the internet. You could build massive habitats before you bothered to send people. Building a magnetic launcher would be trivial with all the moons resources. We could start building all those 1960s style space habitats on whim. Exponential growth is just cool and even scary. <br /><br />Uh oh, the old Von Neumann machine discussion again.<br /><br />Basically, if we had a machine capable of duplicating itself, exponential growth would be obviously possible. No matter how large or complicated this machine was, it wouldn't matter. We would only have to make one of them. So cost wouldn't matter, after the first machine was built.<br /><br />We haven't really got anywhere like making a Von Neumann machine, though many attempts have been made. <br /><br />I read an article here sometime back describing bots composed of cubical parts which could assemble another such bot given all these other parts. However that is not a true Neumann machine, a true one would replicate itself from naturally occuring materials.<br /><br />

 

[thereiwas]

Obviously you have to bring it from Earth, but you should not expect the same techniques to work. When I hear about "robots on the moon" I imagine multi-jointed things with lots of bearnings to go bad, like Japanese industrial robots.<br /><br />I am interested to find out what exactly his new burying technique is that has no moving parts. A rocket engine mounted sideways on a sled that "blows" the regolith over the module?

 

[kelvin_zero]

Thanks for all your replies.. Yes I think you have convinced/reminded me that a human in a workshop is still fantastically more adaptible than any current robot. Great for fixing things that go wrong. Although we use robots for assembly on earth, its easy to forget the number of humans probably required to keep those robots working.<br /><br />Does anyone have any links to scientists estimates of how far off though?<br /><br />Btw.. Im trying to avoid the von newman term, or selfreplicating machine. Im assuming all remote manipulation by human controllers at first. <br /><br />Exponential growth is not science fiction, its just that on earth we hit the bounding limits moderately quickly.<br /><br />If your factories required people on the moon then of course you could still get exponentual growth through the usual means <img src="/images/icons/wink.gif" />

 

[qso1]

Kelvin_Zero:<br />Does anyone have any links to scientists estimates of how far off though?<br /><br />Me:<br />The scientific community really dont speculate too much on time scales of possible future advances. Estimates when they do come about are generally in the 25-50 year range for something such as self replicating robots. There estimates are educated guesses at best. <div class="Discussion_UserSignature"> <p><strong>My borrowed quote for the time being:</strong></p><p><em>There are three kinds of people in life. Those who make it happen, those who watch it happen...and those who do not know what happened.</em></p> </div>

 

[MannyPim]

Greetings Kelvin... interesting question.<br /><br />My own opinion is that rather than self assembled robots, in teh next 2 to 5 decades or so, the predominant technology will be cooperative robot colonies.<br /><br />One of the critical things we ned to do on the Moon, for example, is excavation. When we build true large scale permanent structures, it makes the most sense to build them underground. So how do you excavate large underground tunnels and vaults?<br /><br />The conventional apporach would be to use TBMs (Tunnel Boring Machines) designed for the Lunar environment. However, with a sufficently advanced robot colony, you could do a better job, using much less energy, wiht greater desing flexibility and with almost zero risk of catastrophic failure.<br /><br />Let's say you sent a 5meter diametr TBM to the Moon (which is small by modern standards). Even thought it is small for Earth tunneling, it would still be a monstrous machine, very heavy and very complex and it would only be capable of excavating cylindrical tunnels.<br />Now suppose the machine breaks down like one major component fails.... you're done ! (catastrophic failure mode). You'll have to send anoterh TBM to the Moon, and extricate the damaged one before continuing.<br /><br />If, instead of aTBM, you sent a robotic "ant colony" programmed to excavate a complex underground structure, they would be able to do it virtually unattended. You would send perhaps 500 to 1000 identical robots. Each would run a very simple program (just like individual ants have very rudimentary behaviors), but when working togetherm, they would be capable of quite amazing feats of engineering construction. These robots would not be the size of ants of course. Perhaps more like the size of a cat, let's say...<br /><br />Now, if during the course of building this underground strucutre, some of these robots were damaged or failed, the rest of the colony would continue without a very notceable impact (graceful degradation mode). <div class="Discussion_UserSignature"> <font size="2" color="#0000ff"><em>The only way to know what is possible is to attempt the impossible.</em></font> </div>

 

[docm]

Or, if you're Bigelow, you use inflatables building at L1 then land the base intact, as he mentioned in the Cosmic Log interview, then cover it with regolith. <br /><br />How to do this he explained as a 2 man operation to be tested at his facility in Nevada soon. Whatever he comes up with given his experience as a construction contractor it should be interesting.<br /><br />One interesting method was posted to the New Scientist Space page today;<br /><br />Article.... <br /><br /><blockquote><font class="small">In reply to:</font><hr /><p><b>Magnetic 'elephant trunk' sucks up lunar soil</b><br /> /><br />Future astronauts living on the Moon will need lots of water, oxygen and other resources that can be extracted from the lunar soil, or regolith.<br /><br />But collecting the large quantities needed with front-end loaders and dump trucks could throw up a lot of dust. That could cause a host of problems, since the tiny, jagged dust particles could clog machinery and even harm astronauts' health if inhaled.<br /><br />In a bid to solve the problem, Benjamin Eimer and Lawrence Taylor, both of the University of Tennessee in Knoxville, US, are developing a magnetic device designed to collect soil without creating clouds of the powdery dust. "This idea is akin to a leaf sucker," Eimer told New Scientist.<br /><br /><b>Magnetic coils</b><br /><br />The idea is to build a flexible tube with magnetic coils spaced at regular intervals along its length. Because lunar soil particles contain a lot of iron, the magnetic field produced by the coils would suck the soil into the tube and whisk it along its length.<br /><br />A relatively small tube would be used by an astronaut or robot to pick up soil and feed it into a larger magnetic 'pipeline' leading back to storage facilities or processing plants at the lunar base. Many of these flexible tubes could be attached to the same</p></blockquote> <div class="Discussion_UserSignature"> </div>

 

[MannyPim]

The Bigelow Moon base is a fine concept for an early base.<br />What I am talking about is a permanent large scale base - the first lunar city, let's say - capable of housing upt to 300 or more people and built to last at least 100 years.<br /><br /> <div class="Discussion_UserSignature"> <font size="2" color="#0000ff"><em>The only way to know what is possible is to attempt the impossible.</em></font> </div>

 

[docm]

Then how about this?<br /><br />PDF.... <div class="Discussion_UserSignature"> </div>

 

[kelvin_zero]

Hi,<br />re the post about not speculating on future advances, thats reasonable but I think you might be talking about self reproducing machines. Im talking about factories controlled remotely by human operators.<br /><br />Ive googled the various issues, and keep finding papers claiming to have solutions. If we have all the solutions, regardless of how crappy, then we could put together a non-speculative estimate of what it would take using only these known solutions. <br /><br />The value might be a billion tons, but it wouldnt be speculative.<br /><br />However im also ok with finding out that there are a list of problems that have no current solution, especially if someone can provide a reference.

 

[R1]

I think a robot program of this magnitude could be as costly as the space program itself and distract a lot of funding away from building a big interplanetary spaceship that we need more.<br /><br /><br /><br /><br /><br />The easiest to materialize in the immediate future is probably the current Constellation program<br />using a few lunar landers very efficiently, converting them to modules of a mooncamp.<br />Each mooncamp at whatever site could consist of 6 or more lunar landers and/or their payloads.<br /><br /><br />There's no need to drill tunnels either. <br /><br /> Building factories that build factories that build lunar robots that we don't even know how to build yet<br /> entirely by remote control is even kind of expensive here on earth, and people actually want to go to space and the moon and mars,<br /><br /><br /> and alot of people are even willing to pay tens and tens of thousands of dollars, up to big fractions of a million<br />per each trip to the moon or less, so there's actually less funding if you don't send humans to space.<br /><br /><br /><br /><br /><br />the funding problems of the frequently cut budgets, <br />inflation between design and implementations,<br /> the cost of congress and house and gridlocks , <br />and higher cost of sending artificial people, <br />or building artificial people away from earth is not in the practical future.<br /><br /> <div class="Discussion_UserSignature"> </div>

 

[MannyPim]

<font color="yellow"> the funding problems of the frequently cut budgets, <br />inflation between design and implementations, <br />the cost of congress and house and gridlocks , <br />and higher cost of sending artificial people, <br />or building artificial people away from earth is not in the practical future. <br /></font><br /><br />You are right.<br />As much as I love NASA I have very little confidence that they will actually open up the Moon to large numbers of people. I dare say that I am even doubtfull they will be able to achieve the goals stated under the VSE withing the time line and budget targets announced.<br /><br />So, my hope lies with commercial space and private enterprise.<br />I think people like Bigelow, Jim Benson (SpaceDev) , Gene Meyers (Space Island Group), Elon Musk (SpaceX), Jeff Bezos (Amazon & Blue Origin) will begin to create dramatic advances in space technology that will surprise and astound much of the world. <div class="Discussion_UserSignature"> <font size="2" color="#0000ff"><em>The only way to know what is possible is to attempt the impossible.</em></font> </div>

 

[kelvin_zero]

Hi, I accept what you say is probably true, but I expect this can and has been argued quantitatively rather than qualitatively, by really bright and informed people.<br /><br />Im really looking for existing research rather than to launch an attack on anyones ideology.<br /><br />Anyone got any good links? When I google it I usually find stuff from the 1980s. Presumably NASA has a summary of current status of various in-situ manufactuing technologies that it continually updates.<br /><br />Of course there is the wiki article with several links<br />http://en.wikipedia.org/wiki/In-Situ_Resource_Utilization<br /><br /><br />

 

[R1]

well it's alright, I mean posts discuss various ideas, technologies, problems, etc<br /><br /><br /> to me, populating the moon heavily in this magnitude and without humans as you describe is kind of an unattainable ideology to me, until we <font color="yellow">first</font>implement a good space program and learn the moon, mars, and other moons and places. <br /> <br /><br />Bill Gates and/or Microsoft experimented with robotics, I think, but I haven't heard anything recent.<br />Maybe if someone else knows if Nasa or someone is researching that they will let us know sometime.<br /><br />Could your manufacturing idea be in orbit instead?<br /><br />or must it be on the dusty moon necessarily? <div class="Discussion_UserSignature"> </div>

 

[kelvin_zero]

I have no plan, but the plan I do not have requires the moon <img src="/images/icons/wink.gif" /><br /><br />The fact that in situ resource utilisation is seriously being considered does open up space to a much wider selection of people though. Space is no longer just for rocket scientists.<br /><br />Investigations can be done here on earth with simulated moon dust by anyone with an idea to try out. Computer scientists, geologists and engineers can finally 'have a go' at schemes that could potentially reduce the necessary payloads a hundredfold.<br /><br />No reason why development should necessarily be taken out of NASA's budget, and no reason why they should take up room on any launches until the technologies have been proven.

 

[dragon04]

<font color="yellow">The fact that in situ resource utilisation is seriously being considered does open up space to a much wider selection of people though. Space is no longer just for rocket scientists.</font><br /><br />In current economic terms, in situ is the most realistic approach.<br /><br />I think the key is going to be that we'll have to have more of a "pioneer spirit" as opposed to spending big money and big time training construction workers to be astronauts, or the converse.<br /><br />We're going to need a larger and more "portable" workforce than current convention allows for.<br /><br />IOW, a force of robots are expensive, technologically challenging, require a reliable power source and prone to breaking down.<br /><br />Humans are cheap, adaptive, reliable, and (I hate to use the terms, but if the money is right) "expendable".<br /><br />Humans already do very dangerous jobs that can kill them at any moment in exchange for reasonable compensation.<br /><br /> <div class="Discussion_UserSignature"> <em>"2012.. Year of the Dragon!! Get on the Dragon Wagon!".</em> </div>

 

[kelvin_zero]

Currently humans are far more expensive in space, but I guess you mean <i>given</i> that biospheres can be maintained long term (and perhaps gravity health problems solved). Biospheres are another tech we can research here on earth, which is great. I wish more emphasis was placed on them. It would be nice to think that anything we put on the moon could be expected to survive long term.<br /><br />again though, although I can accept robots are expensive, thats only meaningful until they can be manufactured in situ.<br /><br />again, I accept this is not currently the case nor am I arguing it is close. I dont know and I am looking for research.<br /><br />Ive done some more trawling of the internet, but usually in situ research seems to stop at the chemical equation level. For example I havent seen any designs that take moon dust and bake solar cell tiles, just papers saying it is plausible.<br /><br />Im not sure if this means actual implementation is a long LONG way off, or perhaps preliminary designs are not the sort of thing companies would share on the internet the way scientific papers are.<br /><br />The sorts of research I was hoping to find would have looked something like this:<br />Materials:<li> Aluminium sheets.<li> Ceramic magnets (actually did see a paper on this)<li> Insulated electical wire.<li> glass or other strong nonconductive candidate.<br />...<li> other vital materials I havent thought of<br /><br />Then it would progress to the basic machines:<li> electrical coils<li> low friction barings<li> electric motors<li> a grinder<li> a drill<li> a lathe<li> tool for cutting aluminium<li> tool for bending aluminium into L shape rods<li> tool for bonding aluminium.. arc welding ok in vaccum??<br />...<li> Other vital tools I have forgotten.<br /><br />The above totally omits how these things are constructed. Postulating humans is fine by me. What is interesting is that at a certain level of complexity, that machine shop should have all the tools necessary</li></li></li></li></li></li></li></li></li></li></li></li></li></li></li>

 

[R1]

I don't see a problem with bringing several payloads with tools and equipment. We could probably do that in a few months. I just think the surface base should be realtaively small, more like a lunar camp.<br /><br />It would be interesting to see where the research is so far, too.<br /> <div class="Discussion_UserSignature"> </div>

 

[rpmath]

<font color="yellow">IOW, a force of robots are expensive, technologically challenging, require a reliable power source and prone to breaking down.<br /></font><br />The moon is 1 light second away…<br />Robots don’t need to be expensive intelligent equipment, they can be just remote controlled tools.<br />Put someplace a communication link to earth and control the robots from there.<br />You will need the link to work with humans too, so it is no extra cost.<br />The communication from this point to the robots can be done with cheaper equipment, like those used on Earth.<br /><font color="yellow"><br />Humans are cheap, adaptive, reliable, and (I hate to use the terms, but if the money is right) "expendable".<br /></font><br />Cheap? <br />It isn’t cheap to put a man in the moon and keep him alive... <br />to put a robot is cheaper.<br />Expendable?<br />You are supposed to return humans back to earth, but we can send a robot to do a job and throw away the robot once it is done.<br /><font color="yellow"><br />Humans already do very dangerous jobs that can kill them at any moment in exchange for reasonable compensation.<br /></font><br />It is not the same to say somebody “this is a dangerous work” or to say “after the job is done we wont continue giving you supplies”

 

[pyoko]

I've sen a program about an automated robot that can build buildings (one piece) complete with walls and windows. it does it layer by layer, like a 3d printer from bottom down. The program had a sped-up video of a building being built. Quite impressive. The future usage was not mentioned as being for other planets, however, but for cheap simple accomodation for the poor.<br /><br />http://findarticles.com/p/articles/mi_m0NSY/is_7_22/ai_n6102235 <div class="Discussion_UserSignature"> <p> </p><p> </p><p><span style="color:#ff9900" class="Apple-style-span">-pyoko</span> <span style="color:#333333" class="Apple-style-span">the</span> <span style="color:#339966" class="Apple-style-span">duck </span></p><p><span style="color:#339966" class="Apple-style-span"><span style="color:#808080;font-style:italic" class="Apple-style-span">It is by will alone I set my mind in motion.</span></span></p> </div>