[youtube=http://www.youtube.com/watch?v=cI9H4FoA0b4]
According to the video, Ecce Robot is a new paradigm in robotics. We don’t know if that’s true or not, but we really enjoy the drive system. They have mimicked the biological structures in humans using elastic cables and cheap drill motors as muscles. It is intriguing to watch the complexity that even a simple arm lift requires. This does show inefficient this type of set up is, but we still think it is cool. We don’t understand the desire to use cheap drill motors though. Cost aside, the control problem they mention seems like it could be resolved with a little better motor setup. Then again, we’re sure they thought of that. This seems like a perfect time to bring up a common question. Is it worth the inefficiency of trying to mimic our natural biological structures with hobby robotics pieces? What technology would have to be present to make it worth the complexity?
[via Engadget]
honestly, by now i thought there would be prototypes of a humanoid robot that used clusters of Nitonol Wire to simulate muscles. If you havnt used nitonol wire, you are missing out. A minimal current applied across it causes the wire to contract like a muscle, and just a little amount of force is required to stretch it back out. two bundles could pull on each other, and in effect perform the same mimetic actions as say the muscles in human arm would.
time and time again, over engineering leads to costly end products which are “ok” and “safe” for the most part – but are usually not universally applicable because of high cost. if we extend the cost metaphor to nature we can easily see an over designed biological entity taxing on its surrounding ecological system.
time and time again, cost reduction is absorbed into better software or, more specifically, software that can learn and adapt. what is seen as tremors in ECCE’s movements reminds me so much of a baby trying to walk – it’s obvious the next step is not to over engineer the motors – but to “learn” how to drive the motors you have.
cool
I learned this lesson early in my design career: if you list safety as a feature in a design it instantly calls you out as having nothing else to say. Unless your design was specifically for children or something else where safety is a major concern the fact that you can make something work without maiming people is taken for granted.
Biomimetic structures in robotics are pretty damn cool. It was pretty awesome to see the strands of cable slipping in and out of the creature’s solar plexus. I’m just wondering why you’d go so far as to build a complete torso demonstrating this technique.
I am impressed. My first thought was “holy crap, that could lead to better hybrid types of prosthetics”
And with the polymer? You could make them look more elegant and more life like. Also I am sure that the polymer weighs less than steel and is manufactured proprely, you can make them just as strong.
I think the argument for robots that may be inefficient but look and operate more like human anatomy is for our acceptance of them among us.
If we’re making robots to drive cars for deliveries, or to build cars in factories, then we probably don’t care much about what they look like. But if a robot is to be stationed in our homes and cook for us, clean for us, possibly babysit for us, (or of course the sex/companion robot), then making that robot be as close to a perfect copy of a human as possible may be a required prerequisite for full acceptance.
On the other hand, one might also say that mimicking a human exactly adds a bit of creepiness too though.
I might add that personally, I think the furthest I would want to go would be for something like a cleaning service to have one or two humans supervising a group of robots doing the cleaning, possibly in a couple locations at once.
We have cleaning people come every night in the office where I work and I could completely accept one human supervising 5 robots cleaning throughout the building.
I’m not sure about a robot actually living in my home or leaving any responsibility with one. We’d have to make a TON of progress for that to even cross my mind.
I think we’ll make some great strides, but honestly in my lifetime I don’t think I’ll even have to make that choice.
what ever happened to pnumatic air muscles?
efficiency is primary importance. who care how it looks it not sex toy yet
@googfan,
I’ve seen these used. Unfortunately they usually require a compressor which is cumbersome and loud.
@xeracy,
Nitinol is sloooooow. I have seen robots use it though:
http://www.linuxfocus.org/english/may2001/article205.shtml
@all,
This type of tendon/bone system is common in hands/fingers. personally, the most striking view of this in action was the shoulder.
Nay, the creepiness comes from those normal robots, which have a human-like appearance but, in motion, look like corpses that are shaken by someone. Applying the actual human anatomy to a robot makes it seem more “alive” and therefor less creepy, but in my opinion robots should be designed to look like robots and not like humans, in that way the so-called “Zombie-Effect” can be easily avoided.
@8-[
I think the creepiness factor depends on the use though. If I were to consider a robot caretaker either for the elderly or children (which I must say is a big IF). I would want THAT robot to be more human looking… and something like a DATA-like capability. The robots from I, Robot for example would not be good enough for me for caregiving. Nor would I want there to be any wireless connectivity. smile
The I, Robot level of human like appearance would be good enough for me for service workers and possibly hot dog cart type vending. But it’s just too cold of a style for human connection.
Just like with pets given to the elderly, there are a lot of plastic ones that don’t work well, but the fur covered ones to. The tactile feel of something ‘real’ goes a long way for human connections.
@therian
Yes, we’re not at sex toys yet, but that’s what I was saying about the purpose of the robot. Right now it seems that a lot of robot projects are really a shotgun approach. Meaning that they don’t really have a refined purpose in mind, they are just proof of concepts. Which is great I think, to further the thought process and experiment with different ideas.
When one makes a robot with a specified purpose though, I would agree that nothing matters except efficiency and other technical factors for service robots (i.e. robotic welders in a factory). But if the purpose is human contact in a care giving role or something similar, then I think acceptance of the robot is more important than certain function and design issues.
For human contact, it won’t matter how perfect the robot functions if nobody ‘likes’ to hang out with it.
@archaic0:
Something Data-like would of course be perfect, but is impossible to realise. You would need another Data wired with the first one, just to manage all them little human-like movements like muscle reflexes, senseless movement etc. As long as we can’t create something like a human we shouldn’t try to hide that behind Realskin and loud servos, closing eye-lids.
Take C3PO for example. He looks totally artificial and moves the same way, but he is still charming, because you imediatly know “thats a robot” and not some dug-up half-decayed fella with a battery up his ass. If a servant-robot like c3po fails in some way he was just a robot, but if Nancy the maid starts bumping against the doorframe of the childrens room while her voice is slowly going out of sync, that would be a desaster…
@8-[
I can see your point there. Seeing a robot act strange can be taken in stride while seeing a human act strange is just… strange.
I agree that a ‘real’ DATA isn’t reality, probably in anyone’s lifetime. And even if it was, by the time THAT tech comes around I think it would negate the NEED for such a thing as a life-like robot. If we had that massive computing power and that kind of tech, we’d certainly have replicators and the like so service personnel would probably be un-needed.
We’re probably looking for a solution to a problem that doesn’t exist though. I mean, I don’t think the majority of the world would accept a robot taking care of children (although maybe we’d be ok with them taking care of the elderly). If we did get to that point though, I think it would take several generations to come to pass. It wouldn’t be as fast as my kids accepting it and so it’s OK from then on. And because the hurdle for acceptance of the idea of robots being trusted to be caregivers is so high, it probably doesn’t matter what they look like. The perfect example may come out, but we’d still be scared of it for quite a while. As we should be in my opinion.
I’m completely on board with robotic vendors, maids, truck drivers, etc… robotic driving built into cars on the highway (not cities) too.
The only concern with all that robotic replacement is job loss, but I have to think that we could manage that because those robots will need maintanence, oversight, and construction. So even though those jobs could be lost, new jobs would be created. Probably not an equal number, but I believe that any such conversion from man-power to robot-power can be done well if it’s handled correctly.
Can we as a people be trusted to do it well? that’s a debate that isn’t worth having because history teaches us that we should not be trusted with such a thing. But we should be able to do it. sigh
In any case, I can’t see us really having to figure all this out for at least a couple more generations. They already had ‘working’ robot truck drivers and that went away for a number of reasons un-related to the tech working or not. I just don’t see us as a society being ready for robots to move out of the lab any time soon. Which is a shame on some levels… and maybe a good thing on others…
Yeah you are right there. If one takes into account, how long the Liberation of black people took (yeah, still ongoing, i know) then I wonder how long the acceptance of robots would take…
@8-[
agreed. although I would bet that accepting robots among us could happen faster just because we aren’t starting out with a large bias. Once they’re accepted though, to go from servents to ‘free’ beings and be treated as such… we’d be on the same song, second verse.
I really don’t think we’ll have to cross that bridge though.
I think an honest most likely path for us at least in the next 2-5 generations would be for us to be able to produce a vendor role robot and have it in service possibly at places like toll booths or other remote type outposts as well as vendors of food and goods like airport shops/news stands and the like. I don’t see robots replacing anyone we might have a conversation with now, but those people behind counters that we ignore for the most part, those roles are prime robot fodder.
And I wouldn’t really feel that bad either, because those people don’t want to be there either.
I would hope that we are able to add robots in some roles without destroying our employment rates though.
I can see it now though. Right now where you have a company with 10 people on staff and feel shorthanded, I could see the ‘media’ blaming robots for the ‘loss’ of 5 jobs if that same company added 5 robots where they would not have added 5 humans due to overhead costs like medical insurance and taxes.
Similar to how RIAA counts a lack of growth as a ‘loss’. “we were supposed to increase our profit by 200% this year, but we only increased by 100%. Therefore piracy has stolen 100% from us!”
I’m sure the politics of robot workers would get in the way of intelligent integration regardless of how much logical sense it might make.
I think what most of you guys are referring to as “creepy” is generally the uncanny valley issue.
http://en.wikipedia.org/wiki/uncanny_valley
One, potentially practical, reason to consider building a humanoid robot (besides the previously mentioned sex toy) is that we, as a society, have spent a lot of time modifying out surroundings and designing our technology to match the human form. A humanoid robot would, ideally, be capable of going everywhere and doing anything a normal human can do without special accommodations like those we have to make for modern robots or people in wheel chairs.
@caleb kraft
You are correct. I had forgotten about that term, but that’s exactly what we’re talking about.
The challenge becomes figuring out what roles we really think will ultimately be acceptable or not in the foreseeable future. And whether or not those roles suffer from the valley.
For vendor roles, I would think that as long as it’s clean, us humans would accept any sort of robot even if it does suffer the valley. Because we’re not spending any time with it really, just paying for something and moving on. But a waiter at a restaurant though, since we have to establish some kind of repor with them and most of us generally enjoy the interaction, this role would either have to be so perfectly human as to fool us for the most part, or blatenly robot.
A robotic waiter though, I think would degrade the experience of most diners. Adding a robot to mcdonald’s wouldn’t be bad, but replacing the staff at olive garden or emeril’s would be a bad thing.
@colecoman1982
Very good point. Although I think in almost all cases of ‘integrated’ robots people are trying to make bi-pedal robots with two arms that would work in a human world. Rescue and war robots as the exceptions of course made specifically to go where humans can’t or shouldn’t go.
but one can make a humanoid robot that is not a biological analog with all the same control structures. we can make a robot with two legs and two arms but no elbows for example, or 4 elbows. The thought process here is whether to mimick a human’s layout exactly for the purpose of us accepting it more (or some other purpose?). Would that even be needed? Or is the best aproach to go with whatever design is more technically efficient? In my opinion, I think the role of that particular robot determines the answer to that question.
Another angle… does it make sense to impose our limits on a robot simply for the purpose of mimicking our layout? Ever tried to scratch the middle of your back? Possible but annoying. I don’t think it makes sense to make a robot that has that same limitation. there’s no reason to impose that limit on them when we have the choice to make sure it can reach and work on every part of itself if needed. (we’d certainly want a soldier robot to be able to remove a sticky grenade from it’s back if the case arose, right?)
There are a lot of fine points to this discussion, which is what makes it so facinating, and I’m sure people all over the world in robotic labs are talking about these same things. I’d love to spend a day listening to their discussions.
Nitinol wire is BAD, it needs many amps to contract, and when you stop the current it takes minutes before it cools down and go back to original size. totally unusable
hmm.. id much rather the first and original aproach made by w01f ,
“”I am impressed. My first thought was “holy crap, that could lead to better hybrid types of prosthetics””
i think we shuld spend more timeon improving HUMANS then ROBOTS….
after all the best way to do anything is to do it yourself?….
also the price of such tec would be high unless you can easly mass produce these robots… for a nation at war with streached butgets the high end fight units might be too costly…
your thoughts?
Lots of interesting points being discussed here.
The video was great, loved the detail and the jerky movements I don’t mind at all. The range of movement more than make up for that.
As far as A.I goes, I don’t think the problem is computer power. Already computers are so much faster than the human brain. The problem as I see it is that computers just don’t work like the brain does. It’s a completely different model and no matter how fast computers become, they’ll never mimic a human brain on the development path they are going down.
To get a robot with a brain as good as a humans is going to require a whole new way of looking at computing. The brain is massively parallel. It’s like having millions of computers all working together, each one really quite dumb but taken as a whole, amazing things are realised. We aren’t able to create computers like this yet but were taking early baby steps with more cores.
I’m a little more optimistic about human like robots appearing sooner rather than later and at costs that make mass production possible. We just need to hit that tipping point where the framework is there to build into. Right now there just seems to be a field of ideas and experiments and no real anchors to start creating a framework to realise this. Maybe the first anchored moments will happen soon, maybe in 100 years, I can’t predict that but from what I’ve seen, it seems like robotics is starting to gather some real momentum, and the commercial applications may be what are driving that so hard now.
Anyway, enjoyed reading the comments here almost as much as the video :)
hahaha 1:12 it can also light and smoke a bong just like a real human
I love when you guys talk philosophy, it really gets me happy. : )
The more I look at this the more I love the construction method. It looks like they might be shapelock plastic. I can’t say how many times I’ve spent more time machining brackets for motors and linkages again and again as a design takes shape and develops. This seems like a fantastic way of prototyping.
WAS THAT A BLUE BONG? GREAT FREELOADING ROBOT GIT YER OWN…;)
STUPID ROBOT..ERRR
SMART ROBOT ACTING IN HUMAN LIKE STUPID MANNER.
Really great site that you have built good info thanks.