Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Robotics Technology

Boston Dynamics' Latest Atlas Video Demos a Robot That Can Run, Jump and Now Grab and Throw (techcrunch.com) 52

Posted by BeauHD from the new-set-of-skills dept.
Boston Dynamics released a demo of its humanoid robot Atlas, showing it pick up and deliver a bag of tools to a construction worker. While Atlas could already run and jump over complex terrain, the new hands, or rudimentary grippers, "give the robot new life," reports TechCrunch. From the report: The claw-like gripper consists of one fixed finger and one moving finger. Boston Dynamics says the grippers were designed for heavy lifting tasks and were first demonstrated in a Super Bowl commercial where Atlas held a keg over its head. The videos released today show the grippers picking up construction lumber and a nylon tool bag. Next, the Atlas picks up a 2x8 and places it between two boxes to form a bridge. The Atlas then picks up a bag of tools and dashes over the bridge and through construction scaffolding. But the tool bag needs to go to the second level of the structure -- something Atlas apparently realized and quickly throws the bag a considerable distance. Boston Dynamics describes this final maneuver: 'Atlas' concluding move, an inverted 540-degree, multi-axis flip, adds asymmetry to the robot's movement, making it a much more difficult skill than previously performed parkour." A behind the scenes video describing how Atlas is able to recognize and interact with objects is also available on YouTube.
This discussion has been archived. No new comments can be posted.

Boston Dynamics' Latest Atlas Video Demos a Robot That Can Run, Jump and Now Grab and Throw More

Boston Dynamics' Latest Atlas Video Demos a Robot That Can Run, Jump and Now Grab and Throw

Comments Filter:
  • I never imagined the first generation of Terminators would use javelins, but I suppose they need to start somewhere.

  • Dude could have grabbed his own tools in half the time. And you shouldn't be throwing things, especially big bags of tools, on the job site because its not safe. Useless and dangerous.

    • Re:still useless (Score:5, Interesting)

      by Iamthecheese ( 1264298 ) on Wednesday January 18, 2023 @07:27PM (#63221224)
      Actually throwing a tool bag is useless and dangerous. Being able to throw a tool bag is essential. Anyone who has worked in a construction-like environment is familiar with needing to heft a heavy object just high enough to get it to its objective. That means knowing its weight, its center of mass, what's around it, and so forth. This is done in construction over and over. Granted humans use this skill to make up for a lack of strength which may not apply to the production version of this robot but it's still the culmination of a large and essential skill set.

      • Re: (Score:2)

        by znrt ( 2424692 )

        it all boils down to the question if the robot can do these things on a whim if needed, or if they deliberately coded it to perform these particular tasks in a very specific way and a very specific environment. which would be an impressive but pretty useless feat. one of the speakers even openly admits this, it is an engineering demo. it means that these potential moves are mechanically possible. in no way it means that they have a robot that would be able to spontaneously pull off such moves whenever some

        • Re:still useless (Score:5, Informative)

          by quantaman ( 517394 ) on Wednesday January 18, 2023 @08:32PM (#63221364)

          it all boils down to the question if the robot can do these things on a whim if needed, or if they deliberately coded it to perform these particular tasks in a very specific way and a very specific environment. which would be an impressive but pretty useless feat. one of the speakers even openly admits this, it is an engineering demo. it means that these potential moves are mechanically possible. in no way it means that they have a robot that would be able to spontaneously pull off such moves whenever some situation required it and in a productive way. tbh, in the "'making of' video" they are pretty honest about it (marketing fluff and all).

          I'm pretty sure it was 99.9% scripted. Just the sequence "worker on platform needs tools, ok, first I need to get that board (which I figured out is strong enough to hold me) in order to build a bridge so I can walk into position to place it there", would be an epic breakthrough in AI if the robot figured it out.

          I see the demo as showing off 3 things.

          1) Computer Vision: It needed to visually locate things in environments well enough to execute the script.
          2) Really good balance algorithms (AI driven?) in order to run around and balance (probably the easiest bit).
          3) Really good hardware to pull off the moves.

          • Everything was visually keyed. The tool bag was yellow and black, the frame was a specific green, the board had a high contrast.

          • Re: (Score:2)

            by AmiMoJo ( 196126 )

            The balance stuff is the most impressive.

            The reason why everyone laughed at Musk's robot is that it used the decades old method of balancing, where the robot is constantly balanced on at least one leg at a time. That's why it and similar robots like the Honda Asimo walk in that weird squat way. One foot must always be on the ground and the robots centre of mass must be over it. It has to squat to keep the centre of mass lower, and to allow it to extend one leg while being balanced over the other.

            Boston Dyna

            • I don't think balance is that difficult of an AI problem, it's definitely not trivial, but I remember stuff from years ago people doing fairly nimble walking and balance using genetic algorithms. Of course that was simulation, making a simulation accurate enough to translate to the robot (and/or doing some learning on the robot) is another big layer of difficulty.

              I think the really tough part is building hardware that's responsive, accurate, and tough enough to be able to pull off those moves. That's not ju

          • Re: (Score:2)

            by Ogive17 ( 691899 )
            At :31 into the video, when the robot picks up the tool bag. There appears to be a glitch in the matrix (something looks off with the tool bag).

            If true, it likely means the entire routine (including camera pathing) was pre-programmed.

            Now, the robots movements and agility are still impressive to me but it would indicate the recognition is still completely fake.

          • It's not AI, it's just a lot of hard work understanding how walking happens. See for example: http://underactuated.mit.edu/s... [mit.edu]

            Boston Dynamics has published a lot of papers on it, you can search for them.

            • It's not AI, it's just a lot of hard work understanding how walking happens. See for example: http://underactuated.mit.edu/s... [mit.edu]

              Boston Dynamics has published a lot of papers on it, you can search for them.

              Well now, I stand corrected [venturebeat.com]:

              There is currently no artificial intelligence used for Spot’s walking control over how the robot plans paths and understands the world around it. “We use pretty conventional non-AI techniques in those systems, because we know how to work with those really well and they’re very predictable,” Jackowski said.

    • They needed to illustrate how good their T-800 without being too obvious. This was the most realistic non-lethal scenario they could come up with. They are a military application company, not a creative studio, you have to forgive them.

      • They needed to illustrate how good their T-800 without being too obvious. This was the most realistic non-lethal scenario they could come up with. They are a military application company, not a creative studio, you have to forgive them.

        They haven't been a military application company for many years. First they were acquired by Google, then Hyundai Motor Group.

        There is no path to get Atlas to any sort of actual application, it overheats fairly quickly, so can only be used for short demos, not real work.

    • Nothing is dangerous just as long as you don't have humans anywhere near the site. In 50 years there will be a nice fleet of robots, throwing their tools or bags of tools around on the surface of mars while they build a base of which they will launch their nukes onto those puny humans below

  • We all know where this is going (Score:3, Insightful)

    by Broom Cupboard ( 8588461 ) on Wednesday January 18, 2023 @07:18PM (#63221204)
    There'd be a DoD skunkworks section already. Probably their introduction to the battlefield will be gradual, starting with unarmed tasks, then they'll get "self protection", and that protection will get more and more capable until the pretence is abandoned.
    • Toy drone with a granade strapped to it is obviously the winning battlefield bot combo.

    • Re: (Score:2)

      by AmiMoJo ( 196126 )

      I think they are a long way from being useful on the battlefield.

      For example, machine vision works great in a wearhouse that is fairly clear, free of certain hazards like slippery surfaces, and full of flat surfaces. Out in a literal field, it might find things a bit more difficult.

      They had a robot mule for carrying stuff over rough terrain for years, but it hasn't been adopted. My guess is that it failed to perform well outside of demonstration videos.

    • It will certainly make war pretty weird. Imagine if Russia and the US had these things in Ukraine. It could easily just ends up being the battle of the robot factories, assuming nobody would want to escalate to nukes, where you'd end up with a region devoid of humans where each side is throwing robots at each other until one of them prevails. It would make the pointlessness of it all rather apparent (if that isn't already).

      • I think the pointlessness is pretty self evident, and your ‘robot war’ would simply boil down to which side could prop up their economy the longest.

  • I nominate (Score:2, Funny)

    by ISoldat53 ( 977164 )
    I nominate this thing for president. Couldn't be worse than what we've had lately.

  • More Interestingly... (Score:5, Insightful)

    by Barny ( 103770 ) on Wednesday January 18, 2023 @07:54PM (#63221278) Journal

    I think the far more interesting bit was the snippet about having their software performing what amounts to basic allostasis. Planning momentum and energy requirements are basic fundamentals of human (and some animals) learned skills. The way it is mentioned in the video they seem to actually be working backwards from how humans use it to having their robots do similar.

  • give that robot a crate of hand grenades or a bicycle & paper route

    • give that robot a crate of hand grenades or a bicycle & paper route

      Or a bag of classified documents and a Secret Service-secured beach house ?

      Pick the beach house of your choice as I do not discriminate here.

  • Take that you protesters!

  • Wonderful...a robot that could shoplift and throw away the swag...good training well maybe, for mall cops or Walmart security.

    Note the sarcasm and chuckle...

    JoshK.

  • I for one (Score:3)

    by Lije Baley ( 88936 ) on Wednesday January 18, 2023 @09:42PM (#63221440)

    Welcome our new physical overlords. Serving them is bound be better than social media.

  • When these things get weapons and armour...woah, Nellie!!

  • If they don't have wives that can report domestic abuse, I hear the NFL is interested.

  • "At Boston Dynamics, we're dedicated to producing robots that can CHASE YOU DOWN and KILL YOU!"

Slashdot Top Deals

news: gotcha

Close