h a l f b a k e r yProfessional croissant on closed course. Do not attempt.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
This has nothing to do with <i>how</i> the armor is power assisted, although those pneumatic muscle thingies are cool. In this design the power armor has (basically) analog pressure plates along the inside that register, well, pressure.
Want your power assisted arm to move? just move your arm in
the suit, and the pressure plates will pick it up, translating it to the suit. Also, since the plates are analog, if your suit is super strong, you will still be able to control it relatively well (i.e., not crushing door knobs with your super-strong grasp).
I know that DARPA is working on an "Exoskeleton" concept, but I don't know if anything like this exists yet.
Exoskeleton thingy
http://www.darpa.mi...ust/matdev/ehpa.htm Darpa "EHPA" project [CaptainClapper, Mar 27 2007]
some info
http://www.spectrum.ieee.org/oct05/1901 [jhomrighaus, Mar 28 2007]
Artist impression
http://i80.photobuc...elessbs/dr_doom.jpg [Jinbish, Mar 28 2007]
[link]
|
|
This control system seems fairly obvious, and very familiar. Read _Starship_Troopers_ to see a SF example of armor control. Or read the Darpa link for "A control system that allows the operator to move naturally." |
|
|
I mean, that's the whole point of powered armor. Or was someone expecting to operate the legs with joysticks? [-] |
|
|
I read it years ago -- I don't remember them talking about control of the suits like that... hmm. I was thinking more of a "Armor" by John Steakly - type power suit, although now that I think about it, I guess that's the way THEY did it too :-[ |
|
|
This concept is well known and obvious; made famous by Starship Troopers as baconbrain mentions. |
|
|
What's not so obvious is that it doesn't work--or at least if it works it's a LOT more difficult than it seems. The problem is feedback, which is particularly troublesome for the legs. How do you provide feedback? By pressing against the person's limbs. How do you sense the person's input? Pressure against the person's limbs. Umm... |
|
|
Discriminating between the suit pressing against the person vs the person pressing against the suit is a challenge which hasn't been solved despite decades of trying. In principle, it should be possible by carefully subtracting out external influences and your own feedback pressure. In practice...it's not so easy. |
|
|
A more practical approach may be to use sensor electrodes to detect muscle nerve impulses. This way, there's no confusion between conscious effort and external influences. Even better is a sort of direct neural interface, which is being researched for controlling robotic replacement limbs. |
|
|
As crazy as it sounds, a direct neural interface might be easier to implement than pressure sensor control! |
|
|
[IJK] your premise is really incorrect there. Feedback is experienced as resistance to movement not as pressure. if the suit always trys to be exactly 1mm away from your body then that is your movement control, when it experiences an obstacle it will stop moving, based on how hard it has to try to get away from you the suit knows how hard to push against the object. By applying a variable logarithmic type scaling to the feedback loop the force generated by the suit will increase non linearly with the force imparted. A variety of scales would be available to the operator and could be selected at will, a delicate mode in which the response is 1/2 the force applied for picking up an injured comrade or screaming baby and a 100 to 1 for smashing through walls or throwing cars out of the way. |
|
|
If you are talking about full sensory feedback then you get into a whole set of additional issues but in theory a powered armor suit would be more analogous to a tank than anything else so the soldier inside really doesn't need to know that a bullet hit his shin or that he just bashed through a wall. |
|
|
Jhomrighaus, the problem is that the pressure sensors really can't distinguish between "resistance" and "pressure". It's the same thing as far as the sensor is concerned. |
|
|
Making the suit try to stay exactly 1mm away from the body sounds just great as long as the body and suit are both in freefall. But as soon as an obstacle gets involved--like, for instance, the ground--things get more complicated. |
|
|
Consider the feet. How is the suit supposed to stay exactly 1mm away from the soldier's feet? Maybe if the suit is equipped with superpowerful drills so it can dig underground to the center of the Earth fast enough to stay ahead of the foot... |
|
|
The thing which has prevented powered suit systems from working in the past is this problem. We've successfully implemented powered arms, but not legs. Just make sure not to lean on anything with those arms--picking a movable object is okay, but pressing against a fixed object is a problem. |
|
|
Neural sensors solve the problem by directly distinguishing between the user actively moving his muscles from forces and movement caused by external factors. |
|
|
You both make convincing arguments. I had never considered what would happen if the machine, say, fell to its knees. The suit would interpret the momentum pressing the operator into the suit's legs as a VERY strong conrol input... |
|
|
The issue can be avoided by putting the operator IN freefall, or something like it. |
|
|
Just put the guy (girl, if Samus) in a bubble of water, suspended in the center by a harness. (or somewhere safe, like in a bunker beneath denver, even better) Make the machine mirror his movements. |
|
|
Obviously no force feedback, or the problems listed by IJK will incur. |
|
|
A system of interlaced acuators and sensors would be more than capable of providing both feedback and sensing. I imagine if you will a sheet of exotic bubble wrap. If the bubbles in the sheet were divided into a checkerboard pattern such that the black squares were force sensors and the white squares were feedback actuators then you would be able to serve both functions. If you then apply this to several scenarios you could see how it would work; |
|
|
1) pick up rifle, Force exerted by the soldiers hand is sensed by the sensor net, as the armor contacts the rifle barrel the force actuators begin to apply a counter pressure to the operators hand providing 1 a tactile feedback of the contact and 2 a slight reduction in the amount of force applied to the sensor grid, the operator could then apply more or less force as needed to collect the weapon. |
|
|
2) kick down door, Sensor net monitors both amount of force applied and rate of change, and upon sensing the kick 2 seperate features kick in, 1 the suit augments the power of the suit actuator and 2 deadens the feedback response so the operator is not injured by the feedback reaction when the door is impacted. |
|
|
3) Push Safe from room, As the suit interacts with the safe the force feedback applys reverse presure which increases steadily as the amount of force the suit exerts increases by pushing harder the amount of force exerted by the suit is greater, the safe begins to move. |
|
|
4) Get hit on arm with frying pan, the impact of the pan on the suit causes a the feedback actuators to respond and apply force to the operators arm. since the force exerted originated outside the armor and no corresponding force was relayed from the sensor net no motion is applied to the armor. |
|
|
You also need to keep in mind that the powered armor does not require full body tactile stimulation and sensing, just at the joints, further monitoring of the force feedback on the suits actuators would allow the armor to distinguish between the operators movements and external forces acting on the suit. |
|
|
I was going to ask about the feet, but then I figured out the answer. Just subtract the resting weight of the operator, and measure the increase due to muscle activity. |
|
|
Or have some other part of the suit bear the weight of the operator (internal harness). Still, why ride inside at all? |
|
|
I was thinking that it might be possible to put pressure sensors over the muscles. That's sorta the same as getting electric nerve impulses, but maybe cheaper. |
|
|
That way the suit could tell that the operator was wanting a little more precisely than by putting sensors at places where bones contact, say. A bulging muscle could send a signal into the equivalent motors. |
|
|
Which would be nice for posturing and posing in the suit. A big monster suit pumping up, going into a flex and crouching with all motors screaming might be impressive. |
|
|
I could suggest suspension by your torso/seat. I could randomly insert my 100G survival pod idea, but I like the idea of fitting people with IO devices all the sudden. |
|
|
Making people independantly activate extra limbs without moving muscles is not a far fetched notion. Input can be through analogous input channels. Feedback for limbs touching something can be analogously related as sensory data on non moving parts of the body. A helmet of limb feedback would probably be a popular choice in that the scalp and head is not moving and can be covered with an input had that touches you in a manner that relays information and it will support quadroplegics. |
|
|
I realize we are talking about a whole different learning curve that defeats part of the intuitive nature of the exoskeleton but it is an alternative. |
|
|
I like the idea of suspending the person inside the suit -- but I was mostly talking about *vaguely* man-sized suits (the classic Braveheart "some men are bigger than others" aside) so suspension wouldn't be terribly possible in something that size. In response to GutPunch, I would think that the suit would have no way of distinguishing between input from the user and external input -- i.e. frying pan hits arm, moves armor, newton comes into play and human arm inside the armor hits sensors inside... |
|
|
Since posting, I've re-read the book "Armor" by John Steakly -- I'm thinking like, hundreds of thousands of pressure sensors in the suit, minus the soles of the feet. the suit would be equipped with a sensor that measures which way gravity is pulling, and an... altimeter? -- to distinguish between pressing on the ground and trying to stomp something, as well as when the suit is on its knees and when the suit-wearer is trying to knee something. |
|
|
or maybe a processor to distinguish between motions of walking and stomping, door kicking and rifle grabbing, so the suit will intuitively augment strength under the correct circumstances. |
|
|
With the suit doing all that thinking, where does the driver come in? seems more like riding a horse. |
|
|
I think we have gotten away from the central point though: This idea is, while an interesting topic of discussion, IN NO WAY original. It is described precisely in the novel Armor if nowhere else, which he author even states he has read. |
|
|
This is also discussed at length in the Posleen War novels by John Ringo. Anything of this sort is going to be a custom 1 off unit tailored to the user, the differentiation questions are not really valid as there are numerous ways to address them as discussed in the links. |
|
| |