Modeled After Ants, Teams of Tiny Robots Can Move 2-Ton Car | Hacker News Reader

Modeled After Ants, Teams of Tiny Robots Can Move 2-Ton Car

nytimes.com

171 pointsdejan41 comments

While I was expecting something completely different than what the video showed, it's still pretty damn impressive. Even though this was performed under pretty ideal conditions, it really shows the future potential of what's possible when you have a handful of robots each about a cubic inch able to pull a car. I would like to know how much they were actually pulling though, since it's a bit misleading to say the pulled a 3900lb car. I'd wager it only takes about 75 pounds(give or take) of force. Nonetheless, still impressive.

Edit: just saw it says 200N which another commenter says is about 45 pounds. I'd still say that's pretty sweet for such small trinkets.

I'm familiar enough with the research on the individual robots (before synchronizing them) to know that these robots are pulling between 1000x and 10000x their own weight. That alone is amazing.

Winches normally lift around ~1,000x their weight if you ignore the cable weight. Considering these don't include the cable's weight and are really slow it's far less impressive than you might think.

Granted, the fist time you see a car lifted off the ground by a dinky 1.5 hp motor it does seem crazy.

Well this is roughly one pound of robots using tiny sticky feet. A winch weighs quite a bit and generally connects to something heavy like a tree right? I'm not sure I fully grasp this comparison.

You're thinking specifically of winches mounted on trucks, which is a very narrow application of the device. A winch is just a mechanical setup that uses a gearing system to put a large amount of force on a cable or rope in exchange for turning the gears a relatively large number of rotations. They are used all over the place and don't have to be big at all:

https://en.wikipedia.org/wiki/Winch

Their point was that the winch needs to be mounted to something stable.

The impressive thing (for me, at least) wasn't the "force to weight" ratio, it was the "traction to weight" ratio.

Tractors, bulldozers, etc., are designed to be HEAVY so they can get traction on loose ground. In fact they're designed so the operator can add weight to them in order to get more traction. That these very lightweight objects can get so much friction was truly amazing.

this is probably where "ant" similarity comes into play - the higher number of legs allows for higher traction as for small point of contact traction coefficient may be higher. And thus the synchronicity becomes very important as each one leg can sustain only small traction force.

That was one of the most disappointing robotic demonstrations and videos I've seen in a while. I fail to see the ant correlation. On the surface I don't see anything robotic, just a bunch of light winches operating in parallel (albeit with a nifty sticky footpad). Are they somehow autonomously synced, or just automatically load-limiting?

Welcome to modern university life! The main goal is not good research but good PR to get more money.

The article talks about how they are synched and using the combined force of tiny sticky feet (like ants)...

The researchers observed that the ants get great cooperative force by each using three of their six legs simultaneously.

I'm not getting why this is supposed to be impressive. It just looks like they lined up several tiny motors and used a small gear ratio. There's very little ant-like cooperation.

Edit: large -> small

Does your negative impression spring from a belief that you could have done the same thing if only you had the skill, experience, idea, and ability to execute?

I don't have to be a pro footballer to say that Man Utd are playing like crap this year.

You don't, but only because you have experience seeing footballers playing well to compare against.

What's your point of reference for judging these robots? Unless you're well-versed in the field you probably aren't qualified to say whether or not this is good or bad work. Consequently it's not that unreasonable to accept the article at face value. It's not like you've got anything to lose.

You don't look stupid if you believe a believable story so cynically protecting yourself by saying it's unimpressive compared to, say, the latest Boston Dynamics video, isn't really very fair on the people who did the work.

The researchers’ approach is counterintuitive. Rather than striking powerful blows like a football player making a tackle or a jackhammer, they have focused on synchronizing the smooth application of very tiny forces. The microrobots work in concert, if slowly.
The researchers observed that the ants get great cooperative force by each using three of their six legs simultaneously.

“By considering the dynamics of the team, not just the individual, we are able to build a team of our ‘microTug’ robots that, like ants, are superstrong individually, but then also work together as a team,” said David Christensen, a graduate student who is one of the authors of a research paper describing the feat. The paper will be presented this May at the International Conference on Robotics and Automation in Stockholm.

Well, when you dismiss something as "just [oversimplification intended to demean an accomplishment]" it does not sound very impressive, no. But rest assured it is.

I mean, Philae was just a small probe that barely sent back some signals and grainy photos from a comet. I don't see what's so impressive about that.

What is this a robot for ants? It has to be at least 3 times as large!

I'm curious how much electricity did the robots use to move the car? How much distance over what amount of time? Can we get a video of these robots dragging away a human so that doomsayers can panic more?

They moved a 2-ton car... on wheels, on sealed concrete, with winches...

And it's on the top of HN. Come on guys, really?...

Im curious of application. So, hypothetically, they can infinitely scale the net force required to number of robots. However, the movement speed stays at < meter/hour.

Is the next step ensuring absolute precision? What about terrain with minimal friction to hold on to? How does the "ant" "latch on" to arbitrarilly shaped /sized objects? Is it possible to lift the objects onto wheeled bots to enable accelleration or are we trapped at slow speeds?

Its important to note that currently these robots pull but do not push. So for high precision heavy objects movement I think if space, more specifically, when I want to dock my shuttle at a doace station. Unfortunately, only pulling will cause acceleration that would still need to be offset somehow. So, i am not exactly sure where else to gigantic objects need to be slowly and precisely moved by tiny little robots.

That being said, Im excited to find out

They can't scale infinitely. They can fit finitely many robots close to the target. The further the robots get from the target, the longer and thus heavier the tether they need. The more interesting part of the ant behaviour, not addressed yet here, is how they use each other's bodies as tethers and arrange forces so they add up in the right direction.

INDIAN MAN PULLS LOCOMOTIVE WITH PONY TAIL:

http://abc7news.com/archive/8816949/

They pulled a 2 Ton Car.

From the actual paper [1]

A team of six [bots] pulls with [combined] forces exceeding 200 N.

[1] http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=7407333

that's 20kg if I'm not mistaken

It's roughly the gravitational pull of 20kg at sea level (20.4kg or something). Unit-wise, Newtons are N = kg * m/s²: Force = mass * acceleration, Newton's famous formula.

200N is only 45lbs (ignoring unit goofiness). Makes the whole 2 tons part seem gimmicky.

Here's another story on it, including a video

http://www.gizmodo.com.au/2016/03/watch-100-grams-of-robot-p...

I want to know what sort of future-world all of these naysayers live in that 6 tiny robots, together weighing only slightly more than half my smartphone, pulling any load larger than a bread basket, is not impressive. I know for a fact I could not build such small robots to pull anything over a few pounds. I know because I once tried on a whim (the details don't matter). Did someone invent lassos for ants and ant training regimens while I wasn't looking, or something?

The problem for me is that it was over sold. From the title, I was expecting robotic ants lifting a car and carrying it with the whole "ants can lift 50x their body weight thing". What was shown was interesting but a massive let down from the title hype.

Maybe work on your reading comprehension? The title says move, not lift. What title would you have preferred to avoid your current feelings of personal betrayal?

Personal betrayal? Take this crap to reddit.

You're the one who describe a perfectly reasonable title as a let-down. Maybe you're the one who should go back to Reddit.

"Modeled after Superman, I can move a 2-ton car."

Am I the only one who's mind immediately thought about the ocean's eleven style heist capabilities these ants could provide?

Considering how slow they are you'd need some very clueless guards.

There are too many feats of strengths that pull a car, truck, airplane, heavy thing on wheels,etc. Stay tuned after this commercial as these robots prepare to do the impossible! It makes me kinda tune out the video.

It be cool to see if they had some ant like behaviour, like robots recruiting more robots.

It's also how muscles work, isn't it? A ton of tiny myosin heads pulling in parallel.

The video is a bit misleading: https://www.youtube.com/watch?v=CbTCW9CVHLA

Is it misleading? That little girl is significantly larger and stronger than the individual robots.

Fascinating stuff! But why are so many video clips of this kind accompanied by irrelevant musical backing? Are interested viewers incapable of listening to a commentary without added garbage. I use the word because it is indeed a useless addition to the presentation. Or is it there to test my powers of concentration? I don't think so.

Reply on news.ycombinator.com