One Giant Step for Humanoids

An anonymous reader writes "There are a few robots that do amazing things. Honda's Asimo can walk backward and climb stairs. Sega's idog can dance to music. A tougher nut to crack has been making robots walk like humans. Today, scientists introduce three humanoid striders at the annual AAAS meeting. Unlike other robots that have to power every move, these three save energy by letting gravity do a lot of the work. Like humans, they pick up their feet and just let 'em drop. Engineers say they'll inform the next generations of humanoids and also improve design of robotic prostheses for people. And hey, why not send them to Mars to look for those microbes?"

Comments

sites

[r84x]

Here are the homepages for the biped labs of the three universities represented in the article.

Delft

Cornell

MIT

Re:sites

[Saeed al-Sahaf]

I'm sorry, "One great step for hemroids"? Is this a "geek" afliction?

Prosthetics

[BWJones]

Robotic or semi-robotic prosthesis are going to be more and more in demand because ironically of advances in battlefield armor. Current flak jackets (body armor) and helmets are protecting the vital bits of our soldiers, but often limbs (and necks) are sites of damage from explosions and firearms. Many of these soldiers are undergoing amputations either in Iraq or more commonly in Landstuhl, Germany and coming home with prosthetics of varying sophistication.

There are a couple of interesting recent additions to the Internet that cover these issues. One is an article by Steve Silberman in Wired and the other very interesting site is Stuart Hughes blog. Stuart is a world news producer with the BBC who unfortunately stepped on a landmine covering the Iraq war and now writes fairly frequently about "stumpy" and his prosthetic leg.

Re:Prosthetics

[ZephyrXero]

Kind of ironic if you think about it huh? In movies, you always see scientists working on projects to help people only to have them used by the military in the end...and now we're seeing the opposite ;)

Re:Prosthetics

[BWJones]

We scientists are always making stuff the defense department is interested in. You would absolutely be amazed at the possibilities that people think of for basic science research. In fact, I am going to be meeting with a bunch of DOD folks in a couple of months because they are interested in what we are doing. Not everything the DOD does though is involved with taking of life. There is a considerable interest in battlefield medicine and such. At any rate, this is an aspect of the Bush administrations push to applied as opposed to basic research that troubles me. We should not push basic research to the sidelines because that is where advances start from that yes, even the DOD can take advantage of.

Re:Prosthetics

[Infinityis]

Here I thought you were going to suggest using robots instead of humans on the front lines, and then you talked about just fixing humans with robotic parts. I should think someday we won't have to risk lives so often. We'll have robot wars between countries, and people will get used to the idea of not risking their lives to exercise control over others.

course, that's probably a ridiculous notion...

Re:Prosthetics

[Basehart]

"The U.S. can, and is moving in that direction (highly automated armaments, "smart" bombs, cruise missiles, and now in prototyping, fully automated flying drones), but terrorists still prefer the low-tech approach."

Unfortunately that's why terrorists will have to bring their half of the war into our cities and neighborhoods. As soon as they're unable to spill the blood of their attackers, antagonists, liberators or whatever, they'll start coming after mom and pop on their way to Walmart, Anytown, USA to get even.

If we start getting too remote control on their asses the safest place to be in the USA will be on a military base!

Re:Prosthetics

[DigitumDei]

While I think many of the worlds top armies are definitly trying to move towards that; the possible drawbacks are rather worse than many may assume.

Right now, a leader (lets call him bush for the duration of this example), has to be careful when waging war otherwise he will piss off his country thus ruining his and his parties reelection prospects. With the right application of patriotism a leader can get away with waging war, but its still by no means easy to keep up support when parents children are being killed on a distant battle field.

If the army became completely remote then this political backlash would be greatly reduced since the only loss for that country would be machinery and money.

War could increasingly become the first option rather than the last as the costs become more and more tolerable. This would result in a unbalancing of power that the world has never seen before. With one or two countries completely unafraid to send in the army while most of the smaller countries face a prospect of having to support huge human armies to counter the possibility of hi tech invasions.

Drawbacks would be...

* High civilian/armed forces casualties in lesser countries.
* Even more terroism as all smaller countries realise their armies can no longer protect them.
* The rapid destruction of many economies as countries try to keep up.

As harsh as it may sound, it is good that todays superpowers still have people in their armies. The deaths of those people are what keep politicians in check. Without those deaths, without the political drawbacks they bring, future goverments may make our current goverments of the world look like nobel peace prize winners.

Muscles, perhaps?

[Avyakata]

"...save energy by letting gravity do a lot of the work. Like humans, they pick up their feet and just let 'em drop."

That makes sense, but humans don't really just let their feet "drop." Our steps are actually quite controlled...if we just let gravity pull them down, we'd have pretty heavy footfalls, not to mention an awful lot of shuffling...

Re:Muscles, perhaps?

[Anonymous Coward]

Actually, what you say *seems* intuitively right but is proved not-quite-correct by research in passive-dynamics. The energy required to make precisely controlled steps (i.e., muscularly-actuated) is much greater than the energy humans (or any other legged creature, for that matter) expend on walking. Bipedal walking for animals of our size is possible only because humans have evolved adaptive, energy-saving strategies for bipedal motion (for instance, the long tendons and ligaments of our legs are used as passive energy stores).

However, this is not to say that human walking is not a complex coordination of many muscular systems. It's just that steps are not as controlled as we'd like to think. This is by design, so that we can adapt quickly to unpredictable surfaces. Robots that try to be very controlled in walking usually are very slow because they must do many dynamic calculations that humans simply don't do because of the way our legs are designed.

Re:Muscles, perhaps?

[Dachannien]

Actually, "letting them drop" isn't very far from the truth. Of course, we don't just let our feet fall straight down. We swing our legs forward and let our feet catch ourselves before we fall flat on our faces. We actually let a lot of the motion during swing phase happen via gravity, as our lower legs rotate down and forward around the knee joint.

Probably more to the point of what the blurb was talking about, but didn't really explain: human walking uses dynamic stability. During the period of time where one foot is off the ground, our center of mass is not stably supported by the other foot. Compare this to the insect tripod gait, where at all times the center of mass is within the triangle formed by drawing a line between the three stanced feet - thus making it *statically* stable. And compare this to Asimo and the other famous bipedal robots out of Japan - they maintain a statically stable support by balancing the center of mass directly over one of the legs, but they aren't dynamically stable like humans can be.

Re:Muscles, perhaps?

[cnettel]

Also, remember that a leg is not dropping. I know you indirectly said so by mentioning the energy stores and so on, but I think it's important to make it an explicit point.

We approximate a pendulum rather than letting the foot be some kind of ball attached to a "string" (the leg) bouncing up and down. Human movements without a proper grasp of angular momentum gives strange interpretation, like that of the OP.

Re:Walk like a human?

[bigattichouse]

Its not about efficiency, its about "Creepy" factor. Robots that *look* or *act* human need all those little things that make us feel comfortable... they need to walk naturally, or blink at a normal rate - or you won't interact with them properly, and they give you a feeling of "Wrongness" .

Please.

[mctk]

Oh sure, they can walk like me. But what's their record on Dance Dance Revolution?

Sending to mars it interesting, but...

[WindBourne]

there are probably better forms to send there. The rovers are interesting, but they can not cover a large amount of terrain at a time. It would probably be better to have some sort of a flyer, so that it can move quickly for long distances.

Re:Sending to mars it interesting, but...

[Boronx]

What kind of wingspan would you need to fly on mars? Now, a Titan flyer could get by with little stubs, or maybe some kind of lighter-than-air flyer, that could easily land to do analysis of objects. We really ought to send flyers into the gas giants, too.

Re:Walk like a human?

[gwydion04]

you speak of the uncanny valley, methinks :)

Humanoid! Fetch me a beer! Nice humanoid.

[theworldiswatching]

Humanoid! Fetch me a beer! Nice humanoid.

Why not?

[lawpoop]

"And hey, why not send them to Mars to look for those microbes?"

Probably because there are much more efficient ways to locomote. Bipedalism is risky, especially if you want to bend over a lot to pick things up.

I'm in favor of a radially symmetrical spider-like walker that can turn in any direction, or even invert it legs and continue walking if it gets turned upside down. This would make it much more flexible in navigating the Martian environment.

You could have a central ring with legs attatched all around it, and then a rotating body that includes sensors, power supply, and a grappling hand. The single grappling hand descends from the center and pulls samples up into the body for storage/analysis.

Re:Why not?

[Anonymous Coward]

"Bipedalism is risky, especially if you want to bend over a lot to pick things up.

Speaking as an ex-convict I advise taking this man's words to heart.

Re:Why not?

[Crash McBang]

Well, here's one that was made in the 80s:

Odex 1

Large and strong, this dude could pick up the end of a pickup truck and move it around.

Re:Why not?

[savuporo]

"Bipedalism is risky, especially if you want to bend over a lot to pick things up"

Well, if you want a robust robot it must also be able to get up from any position it might get into. Insects have trouble with getting back on their feet if turned on their backs ( and of course yes they are more stable than bipeds when standing up ). But as you'd have to plan for such robustness anyhow in unknown environments, biped is more practical.
There is another advantage that human-like bipeds hold over other locomotion types - for human environments, like insides of buildings, vehicles, heck - kitchens etc, they are a perfect fit, i.e. the environment is tailor-made for them. For instance, human-like biped, given sufficient processing power and sensory capacity, its a lot easier task to navigate around in a skyscraper, enter and operate the lifts, walk up the stairs etc, than for instance a wheeled robot.

A bit of the old Hubris?

[Exluddite]

Hats off to the folks who are working on these robots. They truly are amazing bits of engineering. But are we really so narcissistic that we think something that looks and acts human is a good design? After all, the robots that really are useful to us (mostly in manufacturing) don't look human.

Mars is gonna be tough

[henrypijames]

It still takes a long way to have those robots learn running, crawling, dodging, rolling like Indiana Jones (or Lora Croft, if you prefer your robots feminine). Until then, I won't recommend them for a mission on another planet.

Seriously, insectoid robots are obviously much more suitable for terrain expedition.

But why....

[CastrTroy]

Humanoid had to be the worse design as far as robots go. For animals it works because it would be hard for there to be an organic life form with wheels. Maybe something like a self driving segway would work well. They have that other segway wheelchair that climbs stairs and everything. If they spent more time designing the robots to do actual task like identifying objects, picking them up, and operating them, instead of spending the time trying to make them walk, we'd be a lot further along.

Re:But why....

[badmicrophone]

well, it's a migration strategy.

If we want our robots to live in a human world in our homes and cities they more or less need to fit our form factor. Additionally, if you want the robots to take over jobs like construction then, at first, they will need to be able to drive the decades old machinery - back-hoes, delivery trucks...you get the picture.

Going past that stage there is also the psycological consideration: a robot with whom you can shake hands is going to garner more emotional investment from us then the tank treaded claw-mobile.

when robots become the new automobile we WILL see them in much more "functional" forms: more machine like and specialized. but there will always be a place for the humanoid robot in our homes.

Great minds elsewhere

[badmicrophone]

I can't help it, but alongside the pride and excitement I feel whenever I see technological progress like this I have this tinge of frustration.

How much money is spent every year on perfume? how many great mechanical engineers are working for sea-doo?

I mean, we could have so much more! Not just in robotics but chemisty, physics, space exploration...

But, alas, I know that all work and no play makes humans a dull animal and that that perfume makes ladies smell very nice. Nevertheless, I cannot help this tinge of disapointment which inevitably follows my rush of happiness.

Re:get on with it already

[Lord_Dweomer]

What an ignorant post. These are generic robotics problems that NEED to be solved if robotics are to evolve into what is today the realm of science fiction. The reason you keep seeing "yet another bi-pedal locomotion technique" is because scientists aren't satisfied with the current offerings.

Rest assured there are entire industries who make specialized robotics for the likes of the fast food industry, etc. What we need is more research into general robotic functions, such as walking, recognition, etc.

Re:Walk like a human?

[Anonymous Coward]

Hey!

Once these nice female-humanoids become more human like, they will stop going out with geeks or even be near them.

Sure they can walk and climb stairs but can they..

[ArmchairGenius]

Do it while chewing gum? I think not!

Re:get on with it already

[scubaed]

But don't invent Yet Another Bi-pedal Locomotion Technique, that problem is solved more than enough to move on to the next problem.

Nope, I'm sorry, but it has barely begun to being solved (and this article is in the right direction).

Asimo does not solve the problem, it merely over-engineers it into oblivion. Linearizing every joint and making it look somewhat realistic does not solve the problem, that's why it can only run 30 minutes or so on a charge (pun intended).

Specialized, but equally amazing I think

[yuckysocks]

This website has a neat video of dancing robots on it. It obviously doesn't carry the same implications
that a low-energy walking robot does, but the motor control and balance gyros and the what-have-you
needed for this act are still pretty impressive.

Video

Source page

Re:get on with it already

[vadim_t]

It is *far* from solved.

There was an article not so long ago about a robot that can stand up from lying on the floor. That was some pretty big progress. However, even that robot is still very far from a human. It needs almost two minutes for that!

Current robots barely walk properly. They still have a long way to go until they can do things like jumping on one leg, which are trivial for humans.

Yeah, Thats Nice...

[St.Anne]

These 'bots are great but they're still kids toys compared to advanced Westinghouse designs from the 1930's. When this robot finishes a task he even takes a smoke break !!

Logical conclusion

[EmbeddedJanitor]

We're somewhere along the following evolutionary parth for war:

One king doesn't like the other king so he goes beats him up. Unfortunately he also gets hurt in the process.

Hey instead of **me** being hurt I'll send some blokes (==soldiers) over to beat up the other king.

The other king doesn't like to be beaten up, so he puts his soldiers in the way. We now have two armies beating eachother up.

Hey let's not send our soldiers into the battlefield to get hurt, let's send robots. Nobody gets hurt. Soldiers can sit at home and eat pizza.

The enemy then gets pissed that these robots beat up their people and build their own. Now we have robots beating up on robots.

Next, the one army gets pissed that their robots are getting beaten up and start hacking the enemy comms to stop the other robots. The enemy responds by hacking the hackers...

So what's the logical conclusion? Is war going to just end up being a big computer simulation with nobody getting hurt? Perhaps the kings should just go decide over a nice game of chess!

bah!

[MistabewM]

A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects. - Robert A. Heinlein

Re:get on with it already

[FinestLittleSpace]

No, it's not. Humans (and animals) have the sense which is more important than ANYTHING in public situations: general awareness.

what if a little kids runs round it's feet, or a kid runs in front of them? The robot goes NEAAAAAARGH and falls over.

Until they can produce organic robots who use their legs without so much bloody automation, then you can start saying theyre ready.

More details in Science research paper, videos

[FleaPlus]

For those of you looking for more details, here's the research paper published in Science (may need institutional subscription) and videos of all three robots.

Here's the abstract text:

Efficient Bipedal Robots Based on Passive-Dynamic Walkers
Steve Collins, Andy Ruina, Russ Tedrake, Martijn Wisse

Passive-dynamic walkers are simple mechanical devices, composed of solid parts connected by joints, that walk stably down a slope. They have no motors or controllers, yet can have remarkably humanlike motions. This suggests that these machines are useful models of human locomotion; however, they cannot walk on level ground. Here we present three robots based on passive-dynamics, with small active power sources substituted for gravity, which can walk on level ground. These robots use less control and less energy than other powered robots, yet walk more naturally, further suggesting the importance of passive-dynamics in human locomotion.

Several different approachs.

[WindBourne]

If you are doing a real flyer, you would need huge wings as the atmospere is thiiiinnnn. But, use some helium/hydrogen in a collapsiable wing, combined with small rockets, it is very doable. Think in terms of a vtol aircraft such as the british harrier. Small wings.

Of course, a small number of ballons with small camera might produce some very intersting results. While we would not have good control over where they went, they would be close enough to the surface to take some very good pix that could be relayed from sat. above.. These ballons could then be landed with small amounts of equipment, but obviously, this is more of a serindipity approach to checking the surface.