Honda’s Dancing Humanoid Robot

http://www.wired.com/2014/04/honda-asimo/

by luu • 12 years ago

65 24 12 years ago

Whenever I watch advanced robotics stuff, particularly humanoid stuff, I'm always struck by feelings that are equal parts awe and mild anxiety.

eksith• 12 years ago

This is normal. If you think about it, at no point in the history of our species has another bipedal, humanoid creature shared our space (excluding perhaps early humans and Neanderthals). We're used to seeing that we're the only things that look like us.

soup10• 12 years ago

No applause haha, tough crowd! I've certainly seen worse dancing. Very impressive demo.

codeonfire• 12 years ago

This is the same stuff from 2011 so they've probably seen the drink pouring, hopping on one foot.

chmike• 12 years ago

The critical factor is the speed and precision of the actuators. And the energy of course. If we could increase the speed the robots could behave much more naturally and do much more.

nebulus2• 12 years ago

It was such a strange mistake for Toyota to abandon their little QRIO robots. Even at their small size, they always seemed to be head and shoulders above ASIMO, and frankly, they still do: http://www.youtube.com/watch?v=fGtxPmcsXfg

(This is an old video. They cancelled their robotics line in 2006.)

tluyben2• 12 years ago

They were Sony robots weren't they?

nebulus2• 12 years ago

Oops! You're most definitely right about that. Toyota is still making robots. Sony isn't.

yukichan2• 12 years ago

Hacker News is turning out to be the RSS feed for Wired's stupid articles and their link bait titles. I've even seen mods change the titles they're so "Upworthy" like. There hasn't been a day there haven't been one or two Wired articles on the front page and I think if they're not gaming it, something I wouldn't put past someone writing obnoxious titles like this, well then maybe we just need a weight against these "Astounding!" Wired articles.

Please just go to http://www.wired.com/ and see one gross and manipulative title after another:

"The Coolest Spaceships Ever Built, Compared by Size" Well I'll only read it if they're the coolest EVER!

"Americans Aren't Ready for the Future Google and Amazon Want to Build" I'm sure they used a scientific poll.

"Why the Console Wars Won't Be Anything Like You Expect" Right, you knew just what I was expecting.

"The Craziest, Coolest Mustangs That Never Were" So crazy, so cool, who would have known?

"Free iPhone 5s Deal Proves Apple Isn't So Special" See there you go, proof Apple isn't special.

Seriously, Wired is shit content. It's a cesspool. Yet every day it's on the HN home page. One obnoxious article after another. Usually HN is a bellwether for what will end up on techmeme or the general tech press the next day, but you think you're going to see these shit articles from Wired on techmeme or the tech page of the New York Times? How often is tptacek leaving amazing comments on these stupid Wired posts. Like never.

DanAndersen• 12 years ago

I'd love to see a machine learning algorithm that can automatically rank headlines on their "Upworthy/clickbaity"-ness. I know there's the extension "Downworthy" that does some text-replacement, but I would really enjoy an extension that could be like a spam filter for headlines.

dang2• 12 years ago

Yes, this was a stupid title (I just edited it), and yes, a lot of Wired posts are fluff and inappropriate for HN. We sometimes penalize sites (without banning them) when they're responsible for too much fluff. But doing so increases the risk of missing any substantial articles.

The last three from wired.com that got significant attention on HN were:

  https://news.ycombinator.com/item?id=7590644
  https://news.ycombinator.com/item?id=7610452
  https://news.ycombinator.com/item?id=7599422

Not the most substantial ever, but not bad.

If anyone wants to throw out some things we might try to address this problem—specifically, the problem of tech press websites that put out substantive pieces but also a stream of dreck—we're interested in addressing this. It's a significant impediment to the quality of HN.

yukichan2• 12 years ago

That's a tough problem. Outside of human moderation I don't have any idea. I'm glad you're aware of the problem though and are working on it. You guys work hard and we pay you nothing!

Someone2• 12 years ago

Very nice demo, but still very far from a useful aid.

- it cannot, I guess, screw on that lid (if it could, they would have demoed it)

- does it work on slightly uneven flooring?

- battery live? Anybody's guess, but older ones (http://en.wikipedia.org/wiki/ASIMO#Specifications) managed an hour or so. This one will not likely not beat two hours.

Also, it still stands and walks like an ape, with bent knees, I guess because the software that keeps the robot upright always needs to be able to bend the knees both ways.

noir_lord• 12 years ago

Stood at the side of Wright Flyer Test in 1903

-------------------------

Very nice demo, but still very far from a useful tool.

-it cannot, I guess, fly 300 people across the Atlantic at 500mph

- does it even work in a 50mph wind?

- flight duration, Anybody's guess but others managed less than two seconds so this one is unlikely to beat 5 minutes.

Also it doesn't fly anything like a bird.

-------------------------

Sometimes things are cool because they are cool and sometimes we need to build useless things to develop the technology we need for cool things.

In 50 years, Asimo and "his" ilk may be regarded as the Wright Flyer's of android robotic evolution.

51Cards• 12 years ago

Regarding some of what you mentioned:

https://www.youtube.com/watch?v=zul8ACjZI18

Someone2• 12 years ago

Thanks! That's a far better way of showing what it can and cannot do. Looks less scripted, and more as a scientific demo. For example, it does not appear to be able to pick up and put down that tray as smoothly as the bottle (did they spend ages to get him pick up and put down that bottle, or are we talking different hardware/software?)

What would top it of for me is a technical description of what Asimo can do, what design choices they made, and what improved between 'versions', if they call them that way. For example, at first sight, it doesn't look like the hardware of the legs changed in years, but that may not be true. Maybe, the thing is half the weight now, the logic for stance control was moved into the knees, a degree of freedom was added, etc.

Theodores2• 12 years ago

Honda should think about using some 'deploy then code' philosophy with ASIMO. I know ASIMO works hard in theme parks and in the lab, however, ASIMO really needs to be in a movie to reach a large appreciative audience. The things ASIMO can't quite do now could be CGI'd in or there could be some (stunted) stunt doubles dressed in a suit.

From what I understand the target audience is elderly care. I have done a fair bit of this myself on a neighbourly basis and I have witnessed how professional care workers can work - they can fill in lots of forms and somehow avoid any actual care. Imagine how dystopian things could get if elderly people had ASIMO and instead of care workers visiting them they just had an engineer to care for the ASIMO. Then imagine the Alzheimer's disease kicking in. ASIMO could be sent off doing the same thing three thousand times a day, with a feedback loop formed, making matters worse.

Clearly the distant relatives would interface with ASIMO rather than their grandma/grandpa. The senile yet devious grandma/pa would do things to outwit ASIMO like regurgitate their pills or eat the bag of sugar instead of their microwave meal. So situations could develop without relatives knowing what was going on. Obviously at things like family weddings the elderly person would not actually go, instead ASIMO would do the duties including standing in the pictures, front-row due to height reasons. You would also have the problem that we have with cars - most houses have lots of them. So some people could end up with more ASIMO than occupants.

So, I look forward to 'ASIMO the movie', where the basic premise is basic care work - as intentioned - and how it works out in practice, ideas fully worked through on the assumption that the code can be written, e.g. to do things like change bandages. Using this 'deploy then code' we could determine if this 'looking after old folk' scenario is realistic.

tempeaster• 12 years ago

This movie already exists, and its great - I've watched it twice. The robot in the movie even looks like the Honda robot.

Movie: Robot & Frank http://www.rottentomatoes.com/m/robot_and_frank/

mwg986• 12 years ago

Was thinking the exact same thing and expecting this movie to show up in the comments. I really liked the movie from a couple different perspectives.

Theodores2• 12 years ago

Thanks for the tip - I watched the trailer and I now have a movie to watch.

blueintegral• 12 years ago

The thing about the ASIMO project is that they do a bunch of research and then publish nothing.

bananas2• 12 years ago

It's a technology demonstrator for a proprietary internal research project. They don't have to publish anything if they're going to use it themselves.

bananas2• 12 years ago

Past the Wired junk, there is some good dexterity progress here. They'll get there one day with the amount of dedication that has gone into the project. I'd kill to work on a project like that.

However I wish they'd put some work into getting rid of the gait that looks (as a human) as if ASIMO has pooped itself.

Honda’s Dancing Humanoid Robot

Related Stories