back to article Scientists unveil a physics-defying curved space robot

A team of scientists from Georgia Tech say they've built a robot that can move without anything to push against – a discovery that seems to violate the law of conservation momentum.  The researchers were able to generate momentum without a surface to push off from by building a robot isolated from outside influences and …

  1. JimmyPage

    er ... doesn't the em-drive claim the same ?

    was my immediate thought

    1. Yet Another Anonymous coward Silver badge

      Re: "didn't indicate the device would actually work. "

      i.e. experimentally demonstrated that the device didn't work.

      1. J P

        Re: "didn't indicate the device would actually work. "

        Absence of positive does not equal presence of negative. I can’t believe it takes a legally trained tax nerd to point this stuff out on a forum where you’d expect people to be quite good on logical thought and stuff.

        1. Yet Another Anonymous coward Silver badge

          Re: "didn't indicate the device would actually work. "

          I think that in the case of the em drive 'Russell's Teapot' applies.

          You invented a machine that does magic, in all the tests it didn't do magic. Claiming that I didn't definitively prove there is no magic is a bit of a stretch.

          1. NeilPost Silver badge

            Re: "didn't indicate the device would actually work. "

            Stop the Magic Steal.

        2. Anonymous Coward
          Anonymous Coward

          Re: "didn't indicate the device would actually work. "

          The ‘em drive design “forgot” that the microwave was in a sealed box. Once you also take into account the angled walls… it all cancels out.

          Other than that you could build a hot air balloon on the same principles…. If you could just keep the air in position…

          When they talk about “curved space” what they mean is that this thing moves due to gravity…. Not as newsworthy really…

    2. Michael Wojcik Silver badge

      Re: er ... doesn't the em-drive claim the same ?

      No; the EmDrive claims linear acceleration with no momentum exchanged with the environment.

      This is a different claim, though I admit reading the abstract at PNAS did not entirely clarify things for me. I get the basic idea of the "noncommutativity of translations" under curvature, but the mechanics as a whole don't quite sit right. (It's worth noting that this paper of course cites earlier ones that describe the theory; this is just apparently the first laboratory demonstration.)

      And the significance statement before the abstract says "akin to how a falling cat can use shape changes to control its orientation but not its position". I'm not seeing that; the process cats use to reorient themselves while falling involves shape changes, sure, but seems to be explained adequately by conservation of angular momentum. (Cats extend their front legs and pull in their hind ones and apply a twisting force in their torso, and that causes their back half to rotate further than their front half; then they reverse the process to rotate their front half. Pretty straightforward.)

      But I Am Not A Physicist.

      1. Justthefacts Silver badge

        Re: er ... doesn't the em-drive claim the same ?

        I *think* what they are saying is this:

        A cat can change its orientation, but not its angular momentum, without exchanging angular momentum with its environment. Because the net angular momentum is zero, despite it being non-zero across a cycle.

        Therefore, an arbitrary object can change its linear position, without needing linear momentum change, by also performing a carefully selected cycle. By performing a repeated set of cycles, the object can move, yet without gaining linear momentum. Their claim is that the secret is spreading the cyclic motion so that the object spans a curved section of space.

        I’m very dubious about this. It smells very much like the (correct, known) fact that you can gain linear momentum by slingshotting past a rotating black hole, known as the Penrose process. This is different to a normal gravitational slingshot, in that you are actually extracting energy from the stress-energy stored in twisted spacetime itself rather than any physical object. Where their claim differs, is that they claim you can effectively extract angular momentum from a *non-rotating* black hole (or other hyperdense object) by rotating yourself across the tidal gradient. That doesn’t necessarily violate conservation of momentum or angular momentum, because it operates via backreaction onto the gravitationally-attracting object.

        However, hand waving argument suggests that the energy isn’t going to come from the black hole, for entropy reasons. Rather, it’s energy that you yourself are expending against the tidal gradients. And if you chose to expend the same energy instead as a collimated beam of light in your chosen direction of travel, you would gain a tiny amount of momentum opposite to the momentum of photons. It’s exactly as efficient a rocket motor as turning on a flashlight in space and hoping it will accelerate you to the stars…..

        I think that what they’ve invented is simply a stunningly inefficient (by 45 orders of magnitude) gravitational wave equivalent of a flashlight version of a solar sail….

      2. ian 22

        Re: er ... doesn't the em-drive claim the same ?

        Are you quite sure about this cat thing? I was under the impression that cats control gravity, much as they control everything else.

  2. Doctor Syntax Silver badge

    If you're that close to a black hole moving without anything to push against would be the least of your worries.

    1. Michael Wojcik Silver badge

      To be honest it's pretty much the least of my worries already.

  3. bertkaye

    Inertia of energy as an anchor in space

    This gets complicated, but all motion in space requires movement in time. That in turn relates to energy transfer from point to point in space. One could then use a large mass as anchor in time to leverage against moving a smaller mass in space and time. Thus a black hole could be used as an anchor for a space drive in the near time vicinity.

    One could also anchor against energy in an E or M field, as the field itself has a kind of inertia in places where it has energy in it. That is, you can use points having energy, within E/M fields, as anchors you can push against, as all E/M waves require fluctuations of energy on a point to point basis propagating through space. In other words, energy itself has inertia and thus can be a stable point to push against. I am trying not to seem handwavish, but this is somewhat complex. Remember that waves carry energy and waves have to propagate through space and time. By carrying energy they in effect can be anchor points to push or pull against. Yes, this implies there can be tractor beams too.

    1. Anonymous Coward
      Anonymous Coward

      Re: Inertia of energy as an anchor in space

      That’s a really complicated and incorrect way of saying “you can build a solar sail”

      Also, it doesn’t explain the tractor beam effect at all (which IIRC is an interference pattern effect using the Electric and Magnetic parts of a wave to attract in the same way any school kid is taught magnets behave)

      1. bertkaye

        Re: Inertia of energy as an anchor in space

        No, a solar sail is quite different. It uses the momentum of applied photons to push it. I'm talking about the fact that energy at a point in space possesses inertia thus provides a place against which one can push. The energy involved in virtual particles constantly emerging from the vacuum then disappearing provides a brief interval when there is an anchor in empty space against which something can push. However these are transient anchors so one can only push briefly on them.

        The downvotes are coming from people lacking a deeper model of physics and they aren't understanding the complexity of the models for MEST. Hopefully eventually they'll gain enough understanding.

        As for tractor beam, I don't mean optical tweezers, I mean that electrogravitic phenomena can synthesize gravity and pull things but also provide the same effect as dark matter, and push things too.

        1. Anonymous Coward
          Anonymous Coward

          Re: Inertia of energy as an anchor in space

          Trust me, the downvotes come from people with a much better Physics education than you.

          1. Anonymous Coward
            Anonymous Coward

            Re: Inertia of energy as an anchor in space

            I don't know. I got a B+ in my GCSE physics and it sounds about right to me.

            (Obviously I kept up that strong academic brilliance after coming out of University with a 2:1 in Fine Art)


            1. Anonymous Coward
              Anonymous Coward

              Re: Inertia of energy as an anchor in space

              No offense to anyone with Fine Art, I just needed to pick something that wasn't Physics haha.

              I would have picked my degree but people might have thought I was serious because we software engineering guys tend to be quite arrogant!

      2. Justthefacts Silver badge

        Re: Inertia of energy as an anchor in space

        At a very deep level, I believe that’s an exact analogy to what they’ve done. They’ve invented the concept of a gravitational-wave equivalent to a solar sail.

        Except that it will be 45 orders of magnitude less efficient, because gravitational waves couple so weakly compared to electromagnetic radiation. And also, there’s no external solar source of gravitational waves, so you have to rely on an internal power source - equivalent to powering your solar sail with an onboard torch,

        1. Anonymous Coward
          Anonymous Coward

          Re: Inertia of energy as an anchor in space

          You mean holding a flashlight out the window won't work? Well, back to the drawing board. Damned coyote and his ACME physics book...

  4. J.G.Harston Silver badge

    That looks just like how I move my wheeled chair across the floor, just swing my legs and varying speeds. In fact, that's how you impart motion to yourself on a swing. Have they just invented the swing?

    1. Schultz

      Ever used a Skateboard?

      Looks like they copied the kickturn (Youtube link).

      No momentum transfer? I call bullshit, because that would break the standard model and, to-date, the standard model has been quite robust. It'll take more than a few students playing with robot parts to overturn that ;).

      If you think you invented mechanical magic, check for momentum transfer via friction.

  5. Zack Mollusc

    Am I missing something?

    From the overhead view, the centre of mass moves toward the 'gate post' when the vertical weights move as far as they can from each other..

    So, if the centre of mass is near the post, when the horizontal masses move to the right and then stop, they will have applied a force on the horizontal arm, moving it left in proportion to the relative mass of the horizontal mass...

    What I am struggling to convey is an idea of the leverage of the horizontal arm.

    Stand on some ice next to a child's roundabou and hold the outside edge of it.t . Have a friend stand near the centre of the roundabout. Now use your arms to slide yourself a quarter of the way round the outside of the roundabout. If this crap analogy works, you will have moved relative to the ground and the roundabout will have moved the opposite way relative to the ground. The amount the roundabout and your friend have rotated will be proportional to the centre of mass of the roundabout and the leverage you get from the radius of the roundabout. Now get your freind to move tt the edge of the roundabout in a straight line so any newtonian opposite force he generates will be directed through the axis of the roundabout and so it will not cause the roundabout to rotate.

    Now, if you pull yourself back around that quarter of the roundabout. Because the mass of your friend is now at the edge of the roundabout , you have less leverage to move him, so the roundabout will move less relative to the ground and you will move further relative to the ground.

    Get your friend to move back to the centre of the roundabout and you, the roundabout and friend are now all back in your relative positions to each other, but the whole apparatus has eotated from it's initial position relative to the floor.

    Sorry if your feet got cold.

  6. Gene Cash Silver badge

    Flat out wrong

    GPS systems, which rely on slight gravity-induced frequency shifts to report locations to satellites

    No. They rely on the time it takes a signal to reach it from the satellites to determine the distance to that satellite, then it triangulates using multiple satellites and their known positions.

    Locations are not reported to satellites, either.

    The whole article sounds like buzzword bingo.

    1. Flocke Kroes Silver badge

      Re: Flat out wrong

      The article just piles wrong upon wrong. Here are the examples I found most egregious:

      the EmDrive uses microwaves in a vacuum chamber to theoretically create thrust

      Early on there was a paper describing how EM Drive was supposed to work. There was a very obvious error in the mathematics: not accounting for the angle when microwaves reflected from an angled surface. The solution to this problem was to delete the paper and not give any explanation for how EM Drive was supposed to work. In theory EM Drive produces zero thrust. In practice any flaw in the experimental design shows up as a microscopic amount of thrust.

      Tests performed at TU Dresden found that thrust reported in initial EmDrive experiments was due to the test unit's interaction with Earth's gravitational field

      Wrong again. The link to the article on TheRegister actually gets it right:

      They found that "magnetic interaction from twisted-pair cables and amplifiers with the Earth’s magnetic field can be a significant error source for EMDrives."

      I just watched the video and it was an omnishambles. As soon as I saw the device I was expecting friction drive: friction is slightly higher when not moving than when there is motion. The usual trick is to accelerate an internal mass quickly enough to overcome static friction and slide a bit then to return the mass with a slower acceleration that does not overcome static friction. This setup is so awful that they barely needed that trick. The vertical shaft is not quite vertical. When switched off static friction is sufficient to keep it from rotating. When only the horizontal masses move their acceleration is sufficient to allow the device to rotate until the unbalanced mass is at the lowest point around the circle. It rotates far enough so we can see the far side of the arm which I suspect used hold a counter weight but was probably removed because the device moves more without it. If there were any 'curved space' magic here the device would be able to go all the way around. If there setup worked at all it should just vibrate without rotating when the vertical masses are stationary.

      To "work" as intended the direction of motion should depend on the phase between the motions of the vertical and horizontal masses. By reversing the phase the motor can reverse direction to overcome some of the complaints normally levelled against fake reactionless drive demonstrations. The "magic" trick here is the vertical masses move on a curved track so they are nearer the bearing when at the ends of the track than when they are near the middle. They are only present to vary how imbalanced the setup is. When the vertical masses are in the middle friction in the bearing is greater than when the vertical masses are at the ends of there track. The horizontal masses then move the arm against low friction in one direction and against higher friction for the return trip.

      I would like to say that as a fraudulent demonstration this one is utter crap but apparently they did manage to fool enough people to get reported here. Please, please tell me this is just click-bait and that no-one really fell for something so dumb.

      1. steelpillow Silver badge

        Re: Flat out wrong

        My favourite:

        Having referenced Relativity, we are told that "objects in curved space should theoretically be able to move without frictional or gravitational forces."

        Except, gravity is curved space(time). These guys offer a contradiction in terms and expect to be taken seriously?

        [Where's the snake oil icon when you want one?]

        1. Fifth Horseman

          Re: Flat out wrong

          That was the big thing that got me too. Please tell me these clowns aren't actually real trained, qualified, experienced physicists.

      2. Zack Mollusc

        Re: Flat out wrong

        Damn it. I like your explanation better. I was taking their word for it re lack of friction. Well done.

      3. Michael Wojcik Silver badge

        Re: Flat out wrong

        Please, please tell me this is just click-bait and that no-one really fell for something so dumb.

        The paper is published in PNAS. I think the problem here is the hash the university department's mar-com department made of it.

        Unfortunately the article itself is behind a paywall (PNAS isn't open-access), but the abstract makes at least a little more sense, and cites prior work published in other real journals.

        I'd suggest that people refrain from posting lengthy denouncements unless and until the read the actual paper, but of course this is the Internet, so that would be ridiculous.

        1. Ken Hagan Gold badge

          Re: Flat out wrong

          I don't think it is fair to make a claim on the internet and then deflect criticism by noting that the *actual* claim was made in a closed access journal.

          If your uni's press office is shit, that isn't my problem.

          1. Anonymous Coward
            Anonymous Coward

            Re: Flat out wrong

            Have you considered Occam's razor here? They claimed something revolutionary, explained it using bunk, then put the evidence behind a paywall. The simple answer is they want to earn some money with their paywall.

    2. phuzz Silver badge

      Re: Flat out wrong

      A better way to phrase it would have been "GPS systems must take into account the slight gravity-induced timing shifts in order to achieve accurate positioning".

      I remember an anecdote (which I can't find anywhere), that one of the original US Military customers for GPS wasn't sure about this new-fangled Relativity thing, and asked for the option to disable the compensation. They found (as predicted) that the clocks on the GPS satellites run about 38ms faster per day than clocks on the Earth's surface, which gives you a compounding position error of about 10km (per day!). The "take Relativity into account" fix was left turned on.

      (This is a combination of the relativistic time dilation due to the satellites moving fast, relative to the ground. This slows down the clocks by 7ms. And the dilation due to being in slightly lower gravity, which makes the clocks run 45ms faster, hence 38ms fast overall.

      And all this is without the very slight changes in signal propagation caused by changes in the atmosphere, perturbations in the satellites' orbits etc. etc. GPS requires a wide spread of different parts of physics)

      A quick back-of-the-envelope calculation tells me that the very first GPS satellite is a whole 0.6 seconds older than a ground spare by now!

  7. Fruit and Nutcase Silver badge

    Closer than you think...

    With the nearest suspected black hole sitting over 3,000 light years from Earth, it'll be a while before we can test that.

    The world is swarming with Black Holes - pick a government of your choice and look at their finances

  8. Kevin McMurtrie Silver badge

    Weird setup

    Less jerky: Two weights spinning at end of the arm on its axis, with the two spinning synchronized in opposite directions to cause cyclical swinging of the arm. Two weights spinning on the axis of the arm close to its pivot, opposite of each other in the same direction so that they only change the angular momentum of the arm.

    This needs three servos moving at a nearly constant speed, other than phase adjustment.

    Not sure if it would prove anything, but it would have less noisy data..

    1. Flocke Kroes Silver badge

      Re: Weird setup

      Back in the darkest corners of you-tube there was a much simpler setup that I cannot find any more. The outer layer is an opaque box to hide the mystical mechanism. Inside is a battery and an electric motor with an eccentric mass attached to its shaft. The mass and motor speed are selected so the magic space drive does not quite vibrate off the table but is sufficient that there is much less friction when the mass is moving back than forward. The magic space drive moves across the table and can be turned around to go back showing that the effect is not caused by a slight slope in the table. With a bit of practice a version can be made that will climb up a clear but gentle slope.

      This new version works on the same principle but is sufficiently complicated that the perpetrators think they can get away with it without hiding the mechanism in a box.

      1. J.G.Harston Silver badge

        Re: Weird setup

        They've invented the badly-loaded washing machine!

      2. PRR Bronze badge

        Re: Weird setup

        > Back in the darkest corners of you-tube there was opaque box ....a battery and an electric motor with an eccentric mass attached to its shaft. The mass and motor speed are selected so the magic space drive does not quite vibrate off the table but is...

        Dean Drive.

        Eccentric weights on a motor, with minimal theory (some say 'a fourth law of motion') and no detailed examination (Pournelle tells a tale). But Dean got John Campbell interested and that got it on his magazine cover:

        Campbell thought it was his mission to expose non-mainstream science for further study. In this case he exposed the stick-slip asymmetry of weight-scale bearings. An interesting story which keeps coming around, as we see here.

        In unrelated news: Dean's Drive In recently closed:

  9. Spherical Cow

    It's just using friction.

    Sit on a swivel office chair with your feet off the floor. Stick your arms out and slowly rotate them left then suddenly jerk them right, you'll spin slightly to the left on the chair. Repeat many times and you can turn yourself all the way around.

    Watching the video it is obvious the robot is doing the same thing. This has nothing to do with curved space and everything to do with overcoming friction at slow vs fast speeds.

    1. Neil Barnes Silver badge

      Re: It's just using friction.

      I find myself agreeing; it looks exactly like there is some not-immediately-obvious friction that has not been accounted for.

      Perhaps they're moving by pulling on a block chain?

      1. steelpillow Silver badge

        Re: It's just using friction.

        I think it's just equal and opposite reaction: accelerate a couple of crawlers the same way, and the vertical-pivoted bit accelerates the other way. Note how the vertical crawlers cancel each other out in this respect.

        1. Snowy Silver badge

          Re: It's just using friction.

          Yes Newton's third law every action has an equal and opposite reaction.

  10. John Smith 19 Gold badge
    Thumb Up

    It's quite pretty, and I like the soundtrack...

    But I'm not sure it does (or shows) what they are saying it does.

    OTOH if it does....

    Smartest thing I know is I don't know everything and I don't know enough physics to offer a quick off-the-cuff.

    My go-to explantion would be "friction,"

    But as always it's when stuff doens'tn work how we expect that actual knowledge gets acuired.


  11. Headley_Grange Silver badge

    3rd Law

    "..without anything to push against - a discovery that seems to violate the law of conservation momentum. "

    Maybe the author believes that jets work because they push against the atmosphere; if so he must be a bit puzzled by how spacecraft manoeuvre in space. As someone above notes, this doesn't look much different to how a swing works.

  12. Paul Kinsler


    Just a word of caution here - do not judge the work merely by reading a remangled press release (even if it *is* on el Reg) and looking at a video. Presumably you might even feel somewhat aggrieved if someone judged your work on such a superficial basis.

    FWIW, there seems to be a pdf of the article on researchgate:

    1. Anonymous Coward
      Anonymous Coward

      Re: But!

      Cool - I’ll wait to read about it in a respected, peer-reviewed journal.

      Free energy nonsense or not, if they have done an experiment that can measure curvature of space that’ll be worth reading about on its own…

      1. Paul Kinsler

        Re: peer-reviewed journal.

        The researchgate "pdf of the article" is, as described, a copy of the published, peer-reviewed PNAS article, which is quite easy to see if you look at it; the link in the Reg article is to the PNAS article page itself.

        Of course, you might perhaps you might have wished to check PNAS publication policy as well, which is fortunately not too difficult to find (nb: two editorial assessments, two independent peer reviewers)

    2. Flocke Kroes Silver badge

      Re: But!

      OK, I wasted time reading that PDF. The first two pages are a Gish gallop of physics mumbo-jumbo. I am amazed that the article here summarising is even remotely coherent. This is standard troll tactic of mashing together so much rubbish that no sane person would even bother to sift through it then claiming that anyone who utterly dismembers even a large part of it did not address the other nonsense that would have made it make sense.

      If you want some fun: by the pictures of their apparatus the description includes "a low friction air bearing and air bushing". At the end of the paper you discover that friction is added back by a polyurethane block pressed against the shaft with an adjustable screw. The device operates by cyclicly varied friction against the support. If they wanted to remove friction they would have hung it by a long thread - and that would have kept it vertical for free without their kinematically adjustable base. They did a version a straight vertical track and got no gross motion because that did not vary the friction by varying the lever arm of the vertical masses.

      I will quadruple down: blatant scam.

      1. Paul Kinsler

        Re: But!

        The relevant (and only) "polyurethane" sentence is:

        "To vary friction to systematically slow the swimmer and probe environmental interaction effects, we attach a block of polyurethane foam to a screw-adjusted sliding block, allowing for adjustable normal contact between the shaft and the foam."

        ... which does not seem to me to be entirely consistent with your hostile characterization. You say "friction is added back" as if it were somehow a deceit, or somehow unnoticed by the authors, but the sentence itself states that it is "to systematically slow the swimmer and probe environmental interaction effects"; i.e. it is part of how they *test* the experiment and its validity. This is a *model* of a theoretical system; and all models, especially experimental ones, are approximate. The relevant questions are whether or not it is a sufficiently good match to their system of interest, and whether confounding effects have been appropriately minimised.

        And the introduction is fine, as academic paper introductions go. Your lack of domain knowledge is your problem, and in no way reflects on the work itself.

  13. UBfusion

    A control experiment

    I would build an experimental control setup using *straight* tracks instead of curved ones, and see what happens. If this behaviour occurs only when the tracks are curved, there might be a possibility that we are in the presence of a new phenomenon.

    I have not read the theory behind the claims, but my hunch is that the control experiment would yield the same result, indicating that there are hidden variables in play, such as friction, asymmetrical air resistance, servo motor & cable inertial effects etc.

    1. Flocke Kroes Silver badge

      Re: A control experiment

      They did do a 'control' experiment with straight vertical tracks and got no gross motion.

      The key feature that makes this friction motor work is that the curved vertical track moves the masses closer and further from the centre. When the vertical masses are near the outside of their track (and horizontal masses are moving say clockwise) they apply less torque than when they are near the middle (and the horizontal masses are moving anticlockwise). Friction against the polyurethane block (described in the paper but I did not see it in the video) is sufficient to reduce the arm's clockwise reaction compared to its anticlockwise reaction to the horizontal masses.

  14. Christoph

    Gaining momentum in a curved space is trivial - just jump off a cliff

  15. The Kraken

    Cats use much the same trick to turn upright when falling, even if initially upside down.

    No net angular momentum, yet they can invert themselves midair without anything to push on - by using intertia of their tail, and limbs in the opposite direction.

    1. Flocke Kroes Silver badge

      They are claiming net angular momentum but they get it using friction between the rotating arm and the support. It is actually in their paper - near the end: search for 'polyurethane' and you will find it.

      This is just a badly disguised, badly optimised slip stick motor surrounded by pseudo-scientific techno-babble.

  16. Fifth Horseman


    In Yorkshire we have a word for this kind of thing. It probably works in other English speaking territories too. The word is "bollocks".

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like