h a l f b a k e r y
Call Ambulance,
Rebuild Kitchen.

meta:

account: browse anonymously, or get an account and write.

 user: pass:
register,

# Self-organising cubes

Organisey.
 (+11) [vote for, against]

[nineteenthly] recently posted an idea involving little cubes acting as cellular automata to form a computer. Which got me thinking about how to make little things organise themselves in space. (His idea allowed the cubes to arrange themselves aribitrarily, and then adapt to their neighbour.)

So.

Magnets are always good.

Imagine a cubic magnet, with opposite faces being north and south poles. Such magnets will tend to stick together in a long row. Not only is this a very dull pattern, but also all of the magnets are identical and interchangeable. Dull dull dull.

Now instead, imagine a small plastic cube with fifteen narrow rod-shaped magnets embedded in it. Five of the magnets run from top to bottom (so, their ends appear on the top and bottom faces of the cube, looking like the spots on the "5" face of a die). Another five run left-to- right (OK, they'll have to be offset a little bit so they miss the top/bottom rods). Another five run front-to-back.

So now we have what looks like a plastic die with five spots on each face.

Now, each of the rods is a magnet (north and south poles at either end, in the usual way). This means that any one face can have several distinct patterns of magnetization:

N_N
_N_
N_N

or

N_N
_S_
N_N

or

S_N
_N_
N_N

etcetera. Obviously, some patterns will be rotationally equivalent to others.

The cube has six faces, making for many many possible combinations of patterns on one cube. Clearly, the pattern on one face will be the reciprocal of the pattern on the opposite face, unless you are cunning and replace some of the rods with L-shaped magnets (achieving, for instance, a south pole on the front face of the cube, with the complementary north pole on the left face).

The point is that there will be a large number of completely distinct (ie, not rotationally equivalent) possible cubes.

Some cubes could also have "null" dots (ie, missing one or more rods, so that some of the "face spots" were undefined).

Now, these cubes will only stick together (face to face) in certain ways. They will only stick strongly when the pattern of north/south on the face of one cube is the mirror of that on the face of the other cube.

Thus, if a large number of such cubes were shaken together (perhaps in water to give them near-neutral bouyancy), they should elf-assemble into some sort of weird structure.

Why? Well, mainly for fun. But if the cubes had some ability to interact via their contacting faces (sort of a bit like [19thly]'s idea), interesting things might happen. Or not.

More generally, the idea is that you can use patterns of magnetic polarity to define many different pairwise interactions, rather like having many different types of magnetic poles.

 — MaxwellBuchanan, Jan 24 2012

Prompted by: Water-soluble_20tablet_20PC
[MaxwellBuchanan, Jan 24 2012]

Prior art: Magnetic_20Dress
Discussion in annos of the same thing, but in 2D [spidermother, Jan 24 2012]

A similar idea... Self-assembling_20rods
...by a similar author. [Wrongfellow, Jan 25 2012]

Super 8 http://en.wikipedia...wiki/Super_8_(film)
Sorry, I mis-remembered the title [Vernon, Jan 25 2012]

self-assembling building blocks http://www.theregis.../02/mit_smart_sand/
not quite here yet [Loris, Apr 03 2012]

cube "robots" with magnets http://www.bofunk.c...75/cube_robots.html
[not_morrison_rm, Apr 03 2012]

Girls and Corpses magazine.. http://girlsandcorp...issues_print06.html
..don't ask me... [not_morrison_rm, Apr 03 2012]

The recent movie "8 millimeter" seemed to include a lot of cubes that were very capable of being organized. This is not so original, therefore....
 — Vernon, Jan 24 2012

Sadly, "8 millimetre" has not come up on my entertainment radar. Can you elucidate?
 — MaxwellBuchanan, Jan 24 2012

 [spidermother] the anno's in your linked idea are generally dealing with ways to arrange identical, bipolar (cue pun) magnets, though I did mention there that a single magnet could have more than one north and one south pole.

The nub of this idea is that patterns of magnetic polarity can give you many different, mutually- exclusive pairings.
 — MaxwellBuchanan, Jan 25 2012

Fair enough. I just thought it was related and worth a mention.
 — spidermother, Jan 25 2012

 If there was some way to instill rules into the cube faces that says, for example:

 If NNNNN then when correct connection (i.e. SSSSS) for x time then turn NSSSN

 This would create disconnect and the cube would go floating through the sea onto the next Face Pairing.

 With this kind of method I could forsee a kind of locomotion to the cube complex. One set of rules might make them crawl, another boil, anther jump, etc..

The problem is how do you get these rods to flip around each other and settle up the new pairings? It sounds pretty stuffy in there. We need a mathematician. Perhaps a rubics cube sort of combination of mini single polar cubes could be used, but you would require intense computation to get them to move into desired recombinations; every single cube needs its own computer.
 — daseva, Jan 25 2012

Nice. But isn't this biochemistry, only bigger, and with magnetic forces substituted for electrostatic ones?
 — mouseposture, Jan 25 2012

I would just like to nominate //Organisey// to be [marked-for-tagline], and suggest the use of hexagons instead of cubes for versitility of curvitude and strength.
 — 2 fries shy of a happy meal, Jan 25 2012

 Somehow, if water pressure could be used to reverse the polarity, the cubes would briefly repel and re-assemble.

By the way, I can see that the cubes can assemble in an overlapped position
 — Ling, Jan 25 2012

 // the use of hexagons instead of cubes //

No.
 — 8th of 7, Jan 25 2012

I'd like this to be done with, not cubes, but truncated icosohedra (i.e. the same shape as the C60 molecule - the same pattern of pentagon and hexagon surfaces arranged in a sphere as a traditional football (US: soccer ball)).
 — hippo, Jan 25 2012

What's the packing fraction for that, though ?
 — 8th of 7, Jan 25 2012

 // isn't this biochemistry, only bigger//

 Exaccerly. More like molecular biology, but yes. Proteins (and other macromolecules, not necessarily the same) will pack together in lots of interesting and specific ways to self-assemble complex structures. That's the sort of thing I'm interested in.

 //the cubes can assemble in an overlapped position//

Good point. It would be preferable if they couldn't. There should be some geometry (of the magnet placement; or of bumps-and-dimples on the faces) which would strongly favour full-face alignments over offset ones, but I don't know.
 — MaxwellBuchanan, Jan 25 2012

I wish my cube would organize itself. Papers everywhere in random stacks with no purpose...
 — RayfordSteele, Jan 25 2012

//He's trying to make friends.// I'd be happy if I could make a simple self-assembling structure. Making a friend sounds far more complex - probably simpler to buy one.
 — MaxwellBuchanan, Jan 25 2012

//Making a friend sounds far more complex - probably simpler to buy one.//
Or at least rent one for a while....
 — gnomethang, Jan 25 2012

There's a theory (which I only understand at a very basic level) that cell development in a growing embryo starts off with all the cells being much the same. The cells then take on different functions depending on where they are in the body. The cells "know" where they are in the body because of the level of concentration of a certain chemical. So, if this chemical is generated at one end of the embryonic blob, that end has a high concentration of the chemical and becomes the head. The other end only sees a low concentration of the chemcal and becomes the tail.

In a similar way, I would like to see these cubes adapt their behaviour (i.e. which sides exhibit which magnetic properties) depending on the local amplitude of a field - this could be chemical (although chemicals tend to be messy) or it could be the volume of sound heard from a louspeaker positioned at one end of the pile of cubes, or it could be an electrical field, or the number of gamma rays received from a small radioactive source, etc.
 — hippo, Jan 26 2012

Your understanding is correct. And yes, the cubes could adapt, if the magnets were electromagnets, so that the polarity of each dot could be switched. But then you'd need quite a lot of power for each cube.
 — MaxwellBuchanan, Jan 26 2012

Just to say i like the idea but i'm a bit busy, otherwise i'd engage more.
 — nineteenthly, Jan 26 2012

[Max-B] I wonder if the cubes could be powered by the same RF field that they're sensing the strength of to gauge their role?
 — hippo, Jan 26 2012

Possible, if you could get enough RF power into a small cube.
 — MaxwellBuchanan, Jan 26 2012

Smartcards and RFID tags rely on RF to power the chip and these are very small - the only really hard bit is antenna design and making the antenna small enough.
 — hippo, Jan 26 2012

I was rather hoping that this idea would follow along the lines of Self-Organising Office Cubes. I had an enjoyable moment reminiscing about Lazy River Offices and thinking about how individual Office Cubes might drift along, form convoys, rendezvous for information sharing and brainstorming with other office cubicles, and dock in marinas for downloading and recharging periods. [All of the polarity specifications could remain the same, provided that they continued to respect the force of river currents.]
 — jurist, Jan 26 2012

 //Smartcards and RFID tags rely on RF to power the chip and these are very small//

Yes, but their power requirements are surely small too. To keep an electromagnet powered up to a significant force would need more oomph.
 — MaxwellBuchanan, Jan 26 2012

 If you want to get some kind of coding to decide whether things should stick together, may I suggest ultrasonics?

 The idea would be to use different frequencies. Similar frequencies would tend to attract (or maybe not repel as much). Different frequencies would tend to bounce off each other more often.

Plus the whole thing should jiggle around like crazy.
 — Ling, Jan 27 2012

//packing// could be interspersed with smaller, more specialized units, like free-floating arms and legs.
 — FlyingToaster, Jan 27 2012

 Saw this video yonks ago (that's an Imperial yonk, not a Metric one), lot of small-ish cuboid "robots" (as they were externally controlled) and they join up with magnets. They also have a kind of diagonal rotation through the centre.

 They aren't forming up to do processing, but if you got hold of some and try the rods and/or replace their magnet with switch-able polarity electro-magnets might save a bit of time.

 — not_morrison_rm, Apr 03 2012

 — MaxwellBuchanan, Apr 03 2012

No extra charge. It still the Rise of the Machines, dontcherknow....<falls asleep with a copy of Girls and Corpses over his face (despite the smell) over his face>
 — not_morrison_rm, Apr 03 2012

To be honest, I think what we really need is self-organising people. Ones that don't bicker, squabble and get in each other's way. Now, that would be an achievement.
 — not_morrison_rm, Apr 05 2012

 I have so many different ideas around this.

 1) Make it a puzzle. 1 cube has a power source, all other cubes are conductive through their magnets. The magnets would not go straight through, but randomly to one of the other 35 possibilities. The power cube lights when a circuit is made would count the number of connections made by measuring the resistance.

 2) Make them spherical with the same magnet arrangement, but place them so that the outer 4 magnets attract and the centre one repels. This would form a grid where the balls do not touch, so you can visually tell if the arrangement is correct, unlike with cubes where you'd need to measure the force required to seperate them.

3) Electromagnetic version. Each cube has a capacitor and can charge another cube. There is a single power source. Any 'flat' cubes would be attracted to all charged cubes. Once charged fully, they would switch on and be either attracted or repelled depending on their arrangement.
 — marklar, Apr 05 2012

 [annotate]

back: main index