Military Robotic Construction Projects

Ok people, I'm just wondering why we cant create a robot that at least has the mobility as the terminator. Ok. So we have Asimo? Ok now get serious. Why cant we have robots that are able to be the same height and dimensions as us, and move like us?

Is it really all to hard to achieve?
There has to be something like this out there in the military somewhere.
I know that if i were a general, and i went and watched Terminator 2, i'd say, Thats it boys, thats our next big budget black project. War machine of the future.

We have movies such as the Terminators, and then we have iRobot. All these machines loooking similar and moving very similar to us. Yet we cant make something that can move like us?

Why not?

Is there any logical reason that we cant?

I mean gods, we have a moving robotic hand in Detroit in the year 2000. And all we have now is a walking robot called asimo who is very short, doesn't look like real human dimensions, and i've seen fall down a flight of stairs. Sure he can run, dance, but only by moving his feet "just" off the ground.


Here's the new iLIMB that has come out in Britain for $17,000 pounds


Asimo falls going up, and also down...


If we have come up with a limb that is able to do that sort of movement especially just for the hand, surely we can get a human sized robot, such as a like of the T1000 ;-), capable of human like movements.

You know what i think? Dont worry about the power requirements yet. Just make a robot that can do this movements and robotics will progress much more rapidly.

You know they want to create robots who can beat human players at a soccer game by something like the year 2016?
I mean cmon! How is it going to happen with the likes of asimo?

iLIMB Thread

Now im not saying make them into a fighting machine of any kind (but im sure eventually many years down the track they could), I'm just saying, shouldn't we have robots able to move like us, that are humonoid looking, and humanoid moving, like us by now? Especially considering our advances in electronics, and technology, all the robotic degree etc out there...???












I just dont seem to understand why we would go every other way into robotics such as robotic snakes, robotic horses, robotic humanoid asimo who's just not human enough, etc...

Thye even had a program that evolves designs to make them better and better such as one to get the best reception via an antenna on a satellite. Why can we get one that will work out the intricasies and nuances of creating better robots with more manuevability?


I'm telling you there MUST be a robot or something somehwere that has at least the manouvability of a human being. And i bet you anything it'd be military. IF it doesn't exist already, it will. We've seen how sci-fi things such as tleporters, have actualll beome real, meaning they have actuallly tleported an object (just a few atoms or so, from one loocation to another)... I think someday we willl have our own humanoid robots, like the terminators ;-P

Does anyone know of any military project or advanced robotics project that contians anything like what im talking about at all?


[edit on 7-3-2008 by DaRAGE]

Good thread with the added videos and examples. What gets me is how most of these companies experimenting with robotics are Japanese based. At these Robotic Conventions you see advancing technology but not much from the US. This makes me think, what is the government working on behind closed doors concerning robotics that we don't know about?

It wouldn't surprise me if we already have some form of humanoid robot that is capable of basic every day human movement. We won't know until the technology we're using surpasses the governments goals. Even at that point we won't know what's really going on because by the time the public finds out, the government will already be working on robotics that is years more advanced than the general publics knowledge. It's the way anything works really.

The US will most likely release a "break through" story of an advancement in robotics that they have kept hush hush for years now. Robotics is a topic that should be more openly discussed because given our track record as a country, we are most likely going to dabble with things we shouldn't and this can create a huge problem when you're dealing with a robot that is capable of superhuman tasks.

Some people laugh but I believe iRobot technology is a possible reality in our near future. I don't want to say iRobot environment because we will most likely destroy ourselves or be taken over before it gets to that extreme.

I can see it now, imagine a war in the hands of robots. Throw some biological/nuclear weapons of mass destruction into the mix with some pissed off robots capable of thinking for themselves. What would you do if you knew you had more power than the people creating you?

Here's a scarey looking robot. Kinda terminator looking but with a hump back

www.pinktentacle.com...

[edit on 7-3-2008 by jfj123]

Originally posted by Shreddy
I can see it now, imagine a war in the hands of robots. Throw some biological/nuclear weapons of mass destruction into the mix with some pissed off robots capable of thinking for themselves. What would you do if you knew you had more power than the people creating you?

Yeah but im not talking about AI that is completely aware and can rebel on its makers. Sure im all for AI as long as they dont want to go on some human killing spree ;-P

But all im really talking about is manouvability. Why dont we have robots / androids that look like us, can move like us? Why do we have robots such as asimo who have completely flat feet, cant balance for shiot, and cant do such things as, stand in one position and spin around on the spot. Is it that difficult? I think not. I think if they had feet, that were capable of moving like ours then they wouldn't have a problem, or hands like ours that can move like ours and do things like our hands do, i mean everything our hands do. Is it insane for me to think these things? Robotics has been going on for a long time now, and considering the amount of university courses, that venture into robotics, it's amazing we dont have something more human like, more manouverable, like us humans. When you think about the long history of robotics, what makes it so hard to have these things? Especially when countries like Japan are increasingly spending large amounts of money on the subject of robotics....

Just a couple of thoughts:

A battle robot is totally different than something that walks around the house and does a maid's job. A wheeled or tracked robot is much more maneuverable due to rough terrain. Even local explosions which would knock a standing robot down, would have to be considered. The complex balancing system for a robot that is knocked down, first to figure out that it has changed it's position while doing nothing, and then calculate how to stand back up; is quite daunting. Now factor in the the machine also has to decide whether to be in defensive or offensive mode, while is happening, your AI comes in here, requires a little computing power.

The standing, human like robot would need as much or more electrical power than the wheeled or tracked. What I base this on is you want a fast robot if it has a high profile. Speed means power, power requires energy. A slow walking robot in a battle situation would last a few seconds at very best. Low profile robots, like in current use, are speedy, but hard to see.

There will be breakthroughs with human like robots. Will Terminator like robots become battle droids, a well place RPG or shoulder fired missile will answer that question. Real life is totally different than a movie.

Humanoid robots are far to difficult to build. For someone who has never actually build and/or programmed a simple robot cannot possibly imagine how difficult it would be to build a robot that can walk upright on TWO legs and at the same time fight in a war.

1.)

The biggest problem humanoid robots (like the terminator) have to face is BALANCE. We humans learn to keep our balance when we begin to walk at a very young age, robots however do not have the ability to learn to keep balance they must be programmed to do so and so far, no one really understands how we humans keep our balance in the first place.

2.)

Another problem is walking around without colliding with objects and effectivly planning which way to go etc. We have a 3d perception thanks to our brain. When we see something in front of us we can estimate how far it is away and how we can go around it. We percieve our world as a 3 Dimensional space (cause that's what it is simplified). Robots cannot see in 3d, all they see is 2D image from the camera mounted on their head which gives them almost no information on their surroundings what so ever so they need all kinds of sensors and whatnot to pevent them from crashing into something.

3.)

Ai is also important, espescially if the robot is designed to defend a country . The robot needs to be able to see the battlefield, to distinguish between friends and enemies, to aim a gun (or something similar) and to make hundreds of decisions in seconds.
So far no one has created a software that can do all that, and even if there was a software like that we would not have a computer good enough to even closely simulate the power of our brain. The biggest and most powerfull computer of our time(2008) was built by IBM and fills up a whole warehouse and it doesn't even come close to the brain power each and everyone of us has. Sure, it can do all kinds of neat physics calculations (about 1 quadrillion / sec.) but it's just not a match for our brains.

4.)

Also a huge problem to consider before building a war machine terminator-style would be on how to make it move. As stupid as it sounds, this is actually a huge problems. There are two main ways to make a robot move nowadays, the first beeing with the help of servos and the second beeing with artificial muscles.

Servos do not provide the speed for fast movement and it doesnt look fluid at all. A robot arm which uses servos cannot move it's arm up and down as fast as a human could. So if a robot needs to react quickly to something (maybe someone firing at it) servos are not the way to go.

The second method is using artificial muscles. These are basically plastic tubes which are filled with air which makes them contract like a human muscle. While these muscles provide a much smoother more realistic and quick movement, they are far to difficult to use for more complex applications.
A humand beeing has over 600 muscles , with each of them made of hundreds of smaller strings etc. To build a system like that artificially with air-muscles....impossible.
There is a reason why it took millions of years of evolution to become what we are now. We are extremly complex beeings and trying to copy that in a few years.....hm...I don't think so (although we're very close).

5.)

Cost and usabillity. Sony build a robot called "Qrio" . Qrio is about the size of and infant, can walk on two legs (very slow) and can have basic conversations with humans. It sold for $50,000. Now imagene building a robot as complex and intelligent as a human..we're talking $1,000,000 per robot here atleast.
And why would you send a robot to war, it's not like its any better than a human...one shot to the head and it's dead....it's not like they would have a metal skin which would be bullet proof...the shock of the bullet alone would kill all internal hardware instantly.

The only reason that I see why we should send robots to war is, that no human has to die in a war ever again.....

sorry for the long post

Sorry for the typos etc. (haven't slept at all)

Thank you for your imput.

Look im really not talking about sending a robot to war (though i still do not doubt that i could happen inthe future somewhere.... ;-P)

But all i am really alking about is the manouvability of a robot.
Really, they want to get robotics up to a stage where robots and humans can play soccer and the robots could win. Wouldn't the best way for that to occur would be to make them humanoid (first off so they can actually play soccer), and be able to manouver like humans.

If you watch the video's we are able to create robotic hands that are very human movement like. Very manouvable and flexible. Sure they have a bit of work to go, and could do with some movement in the wrist and forearm area. But aren't the hands and feet meant to be the most complex parts of a human body? With all the ligaments, small bones, small mucles, tendons. If we can practically replicate the movements of the hand, surely we can get replcate the manouvability of the feet. The toes are a major part of the foot, especially for balancing...If they can get the fingers moving right, i think they'll be able to feet the feet moving ok...

The move i think about balance, the more i think its easier to do than you would think. Just look at the Sony PS3. It has a six axis controller that detects what angle the controller is on. Have a look at the Nintendo Wii with the Wii-mote controller. Hand held, knows what angle its on, how fast it's moving, in what direction it moving. Surely if we put these sorts of sensors around the robot, and one section of the robot feels like it's a bit off centre or off balance, then it could be programmed to correct its self.

About those camera's that can apparently see things in 2D. There are actually programmers who have made cameras that see in 2d, its prcesses the information and creates a 3d image. You can always get three or two cameras situated apart from each other to create a 3d image also...

2D camera that can process in 3d to measure depth...
2d can become 3d and get depth link

I've also watched some doco's where they convert 2d faces of people into 3d images...

Here's the ps3 eye-toy and what it can do. Infrared-imaging buddy.



ps3 eye-toy 3d animation motion tracking
What do you think?

[edit on 10-3-2008 by DaRAGE]

The things in that measure the movement of the ps3 controller are called gyroscopes. They are used in sony's QRIO robot (and you can easily see how unsecure it is when walking at a fast pace). If you were to use them in a human sized robot you'd need hundreds and with each of them atleast the size of a ping pong ball it's not possible.

Image a robot that is standing straight on one leg and has the other pulled straight up. the gyroscope in the robot would sense that it's not out of balance because it stands straight but if the robot was to fall over at that moment (maybe due to wind) it would fall down because the gyroscope does not have enough time to react and send a signal.

As for it beeing easy for us to keep balance, well of course it is, we've been doing it for years now. We just do it (we don't use gyroscopes) and exactly how we do it, well...we don't know. So to keep balance for us is easy , our bodies have changed and we have the ability to learn, a robot however can only follow a fixed set of rules which we programmed him to do, so if we don't even know how WE keep balance, we cannot teach a robot how to do it.

And about the 3d Infrared-Vision. I will not be able to create a 3d image, just the outline by monitoring movement in the infrared light that everyone of us sends out, kinda like the EyeToy for the Playstation2. Although the eyeToy only measures change & movement of the normal light, this would not be good enough for robots.

Important Note: An infrared sensor would not help a robot to sense the difference in the ground makeup, meaning it doesn't know wether its walking on sand,gravel,wood..or even water because it just would see the difference. I give a robot 20mins. with that kind of vision before it breakes.


An most important of all, those infra red light sensors have limits. They might be able to recognise the outlines of a person thats perhabs 5 metres away, that's it..not a person who maybe is 20 metres away.

And what if there's fog or worse, if it rains, the infrared light of the raindrops would mess up the infrared sensor making it impossible to make out any human shape, and while we're at it, how would a robot that only sees theinfrared-light of a person distuingish between a friend and an enemy?

Believe me, I know what I'm talking about, I've worked for RoboWatch (Security Robot Company) in Germany.

[edit on 10-3-2008 by redshirt0202]

Ok then so scratch the infra-red camera.

That doesn't meant we cant have a number of other different camera's like the"Stanford camera chip can see in 3D ". How about all the sensors that those offroad robot controlled cars use for the fema challenge?

Ok so about balance, how about trying to tie up a robot with a human, to learn and remember how to walk, jog or even run properly and stay upright. Get the robot to remember how to do it properly by being joined to a human hook up in hi-tech gear. Like this Sarcoman robot...

Sarcoman Robot Dancing


Sarcoman Robot Playing air hockey


sarcoman robot juggling


SARCO BALANCING!!!!!!


Sarco's Arm Balancing a pole!



Then we have the Exo-skeletal half robot half machine contraption that has as you casn see a lot of flexibility, and agility, and can lift a lot of weight. If we could get a robot to remember how to move and walk with a human in it, then why not without it?

We have to start somewhere with getting a robot to balance and manouver properly. Start with a flat surface... and get it to remember human movements and balancing traits....

I do not see a reason why we would want military robots to look like humans. I guess the human look is only important for social interation , so it will be used to domestic needs. Probably the military would go in the future for Swarm Intelligence robots that would be generated using evolutive computing in order to make sure they are 100% adapted to the enviroment they will have to evolve in.

Alright so I looked at the chip that can see in 3d, although "3D" doesn't really fit here. The chip is able to see depth of field to a certain extend in a 2d still image, but it's not true 3d vision, as it doesn't give the robot a "mental" or in this case "virtual" 3d representation (or map) of it's surroundings. In the future, that chip might be able to do just that but it still has a long (and I mean LONG) way to go.

As for the FEMA challenge, some robots died after 10 minutes and all they had to do is stay on the road. I think that tells us a lot about what robots can actually do once they're send out into the wilderness, I mean, 10 minutes...... and they didn't have to keep balance or fight against rebel armies....they only had to stay on a road twice their width.

One robot (don't remember which team...I think it was the VW) just stopped around a corner because it's infrared camera could not "see" or identify the area for some reason. (this is a bad thing if you're in the middle of a war).

(I don't know if this was also the FEMA challenge bit there was another race where most of the robots didn't even make it past the starting line)

Anyway, all what those infrared-cams did was to distuingish between flat and rough areas. The robot was programmed to stay on the flat area (the road) it didn't actually see it's surroundings. What if the robot had to go trough the wilderness without roads? Not a chance in hell with infrared-cams.

!No way José.!

An infrared-cam or the 3d-vision chip from standford still cannot distuingish between enemies and friends . (This is very important in a war)

You also wrote about "teaching" a robot to remember movements by hooking it up to a human. I'm surprised you didn't see the nonsense in that yourself (no pun intended)

If you teach a robot how to walk on..let's say a small pit of sand. To walk over a pit of sand without falling the sand it walks on would have to be identical to the one it was taught to walk on...if you were to take a hand full of sand out of that pit, the robot couldn't walk on it anymore.

My point is that the area he walks on would have to be IDENTICAL to the one he learned to walk on....the chances that this is the case are extremly small (about 1 : 1000000000000000000000000000000000000000000000000000)

What if he has to walk on ice? What if the ice breaks and his foot is stuck?
What if he slips?

And the most important thing that a robot will never be able to achieve, is intuition. You can't learn it, you're born with it. A robot will never be able to behave and make decisions based on intuition and experience like a human would.

Unless of course we'd find a way to create true artificial intelligence that would be aware of itself...but then we couldn't just tell it to go to war and do this and that..we'd have to treat it like a human beeing.

cheers,

redshirt







[edit on 11-3-2008 by redshirt0202]

[edit on 11-3-2008 by redshirt0202]

Great post with lots of interesting conjecture!

When I got the March 2008 issue of Popular Mechanics, I was treated to a picture of the US Army's SWORDS/Talon robot. I also read a great article, but don't know if it's yet online.

Here is the SWORDS/Talon, as well as variants:
www.foster-miller.com...

The Lockheed Martin MULE was also featured, though it is far from combat capable.
www.popularmechanics.com...

The article states the Army is handling their UGV (Unamanned Ground Vehicle) program with "kid gloves", and the robots fielded will have to get human authorization to fire a shot.

I feel that this article shows the current trend towards tracked or wheeled vehicles as the weapons platform of choice for military application.

Enjoy!

Well they have experimented a little bit. Something like this is being developed but it's as far as I know still in it's testing stage. Least this thing doesn't fall down stairs.



I was more thinking along the lines of a battlemech myself for the military instead of terminator type robots. You know something that can destroy a tank in one volly or lay waste to an entire city by itself. But I see what your thinking.

I don't know what the black box groups have in the backyard right at the moment But I am sure that if something cool comes out of it we will eventually know about it and then of course we will all hate it and start screaming "they took our jobs"



So kinda makes you think if they can make a robot that can do anything a human can what would they need humans at workplaces for anymore?

Ungh...

There are so many reasons. Far too many to post all of them on a single thread.

But I'll cover a few...

Power consumption.
The humanoid design is inherently horribly inefficient.

Accuracy, vs. Speed.
The more you gear a servomotor for accuracy, the slower the limb/wheel will move, the faster you gear it, the less accuracy it has.

Power storage.
Let's face it, even if you find a nice balance for your servo's, you're still requiring allot of power... where are you going to store it? The rest of the body is now occupied with servomotors/gears/solenoids/frames etc. Where is that massive power storage going to go?

Bad design.
The human body, whether we like to admit it or not, is probably one of the most inefficient models to build a machine around.


I so want to get into each area in detail, draw drawings, and show everyone why it's counter-productive to build machines in this manor, but as I said, there are simply too many issues to cover in a thread. I'd have to make a whole website devoted to educating people in order to deliver a single point.

It simply isn't efficient or productive.

Just to bring to the attention of thread readers of some new technologies that could be used...

Scientists To Develop 3D Camera With 12,616 Lenses
Scientist to Develop 3D Camera with 12,616 Lenses

"Application areas for this technology, if it can be successfully developed, include facial recognition, biological imaging, 3D printing or 3D modeling of buildings. Abbas El Gamal's research group thinks that this technology in fact would be superior to the visions of humans, providing robots much better spatial vision capability - "to perform delicate tasks now beyond their abilities""
"The big question of course is: How much will such a chip cost, especially if it has thousands of lenses? Surprisingly, the researchers believe it could be very affordable. In fact, the chip may cost less "than existing digital cameras"."

Self Repairing Artifical Muscle that can charge your ipod

"We've made an artificial muscle that, when you apply electricity to it, expands more than 200 percent”.

The way it works, is that as the artificial muscle material contracts after expanding, the carbon nanotubes rearrange themselves, causing a small electric current to generate, which can then be captured and stored in a battery.

It is also robust. If any part of the carbon nanotube packs up, the area around it seals itself off by becoming non-conductive, thereby stopping the defect from affecting other areas. The clever muscle is also eco friendly and energy efficient, conserving up to 70 per cent of energy pumped into it. Some boffins have suggested that it could even be used to capture wind or wave energy in the future to charge electrical devices.

The research could also lead to the creation of moving robots, better prosthetics, and battery charging energy sources. A self-repairing Ipod would be nice. "

Now in reply to Johnsky:

Power Consumption: The humanoid design with these carbon nanotubes artificial muscles can actually move and generate power. This makes it much more efficient does it not, and wouldn't it take away that big power storage device also and the hassle of where it's going to go?

So the accuracy vs speed: Doesn't the servo get taken away and replaced by the artificial muscles? Shouldn't these artificial muscles react with speed and accuracy?

And hey with this new 3d lens, it should be able to see perfectly ;-P

reply to post by whatukno


Just letting you know that that big dog / small horse robot is a DARPA project. That thing has some excelllent balance....

Here's an updated video of the Big Dog... Just released in March 2008




Now also made by the same people who made big dod is this small little critter who can go through scrub, bush, water, and can even swim underwater....




Hell maybe for now we could get the big dog for balance, and try to put a human like half on top, such as from waste up... with robotic hands, head, etc...

[edit on 24-3-2008 by DaRAGE]

Well DaRAGE, don't let the naysayers of redshirt or johnsky dampen your enthusiasm for this subject. Its obvious they aren't aware of the slew of technologies that exist out there. More than likely, all the juicy projects are not even made public because, honestly, they are easy to copy, so they're kept under wraps. Need proof? Just look at all those copy cat kits for hobbyists in Japan that perform like Asimo. They just looked at the design and said hey, we can make it simply with these common parts.
Billion dollar idea stolen for nothing.
Those robot challenges out there are more about getting people involved in robotics that wouldn't normally do so, and thats where your best ideas come from. These people are not schooled in what you can and cannot do, they'll just pursue a solution until they get it. In fact, the "know it all can't be done and why do it anyway" attitude just demonstrated by johnsky and redshirt, would get them let go from any respectable high tech company pursuing such projects. I know my company would dump them in a heart beat for such defeatist criticism. I know a few companies that regularly profile they're employees and weed them out over thinking like that.
I'm working on 3 patents for robotic muscle systems. All 3 would turn your head inside out on their design. Here's a hint on one: the simplest of the 3 could keep US garment workers employed for quite a long time.

reply to post by DaRAGE


Well, there's always this:

www.youtube.com...

Granted, its not exactly autonomous...yet, although they do hint that the idea has crossed the mind of the designers. So at the very least, the idea is (sort of) being considered.