Lego Robot Solves Bigger and Harder Rubik's Cubes

kkleiner writes "It was only two months ago that we saw Mike Dobson's Cube Stormer Lego robot that could solve any 3x3 Rubik's cube in less than 12 seconds. You would think that there was only one person in the world crazy enough and talented enough to pull this off, but now we have found someone else that is just as amazing. The latest Rubik's cube-solving Lego monstrosity is called the MultiCuber, and although it's constructed out of nothing but Mindstorms components and a laptop, it can solve 2×2, 3×3, 4×4, and 5×5 cubes all in the same build! As if that weren't enough, a larger version solves the dreaded 6×6 Rubik's. We discovered the MultiCuber when its creator, David Gilday (IAssemble), wrote us an email to brag about its puzzle-solving might. Consider us impressed, sir."

heeey....

[socceroos]

That is kinda cool. Can it make me a sandwich? No? Oh.... Well then.

Re:

[Tablizer]

Can it make me a sandwich?

No kidding. The me-too-ism is too much in this case. There are plenty of useful and novel ideas out there. Even a Lego Roomba-like vacuum would be interesting because it hasn't been done in Legos yet, and the technology may kick off other Lego robot servant crazes. The talent could be put to better use than Rubiks++ in other words.

Re:heeey....

[shentino]

http://xkcd.com/149/ [xkcd.com]

Re:

[socceroos]

Yeah, I saw that one coming...

Re:

[Pence128]

if he were running windows, just take out the second and third lines.

Re:

[Tablizer]

...welcome our Lego overlords.

Be careful what you wish for. It could have far-reaching impact on our entertainment [screaming-penguin.com].

It's amazing

[tpstigers]

Just when you think they've peaked, Legos get even cooler.

Re:

[Anonymous Coward]

How much cooler? Does the coolness increase linearly with the size of the cube - 2, 3, 4, 5... n Fonzies? Or does it go as the cube of the size? Perhaps it's exactly as cool as before because it solves the same problem.

Re:

[Anonymous Coward]

No actually most of us stopped caring after the first robot was made. "OH MY GOD IT CAN SOLVE 36x36 cubes" *snores*

Re:

[Enigma23]

Can it solve a Rubik’s Dodecahedron [genomicon.com] yet?

I gues they'll need a Beowulf Cluster to solve one of those... ;p

Re:

[dintech]

If you think that's good, take a look at the rest of the internet. It's awe inspiring.

Re:

[MyLongNickName]

Actually, the more I see people whine about this, the more it makes me want to say 'legos'.

Conditions of cube?

[treeves]

Does it have to be one of those well-broken-in, lubricated cubes that easily spin, or does it work with a stiff cube just out of the packaging? I'd bet it would not.

Re:Conditions of cube?

[SuperMonkeyCube]

The reason that most speedcubers use the Fridrich method is that it is affected less by cube conditions than the corners first method used by Minh Thai in the 80's. A corners first method involves a lot of slice moves (turning the middle layer between the two outer layers), effectively doubling the frictional force required to turn it. The Fridrich method has a large move table and emphasis on face turns instead of slices. The Kociemba algorithm, like what you would find in the 'Cube Explorer' program, uses a much larger move table even than the Fridrich algorithm, and is optimized for a low number of face turns - although not always minimal, it's usually pretty close. So even if the thing wasn't lubed that well, it's going to be fast just because of the low number of moves that it will be able to compute. I would also hope that the robot would be able to apply more torque more precisely than a human and isn't doing more than one cube turn with a given motion, so it has a good chance of overcoming the friction of a new cube. Human cubers do things like RU' (right face clockwise, top face counterclockwise) with a single hand motion and an unlubed cube would hang up on the transition between the two motions. The robot would do two separate clean twists without having to worry about a transition.

All of the V-Cubes, which would be any 6x6x6 or 7x7x7 available that I know of, are more speedcubing friendly right out of the box, as its design was done with correcting for small misalignments in mind so as not to put too much torque on the pieces when turning the cube.

Re:

[DJCouchyCouch]

Flamebait!? I was serious! I was having a true blue nerdgasm.

Sheesh!

Re:

[Sulphur]

Lubed Cubes? Oh well, if it is sensitive to initial conditions then.

Re:Conditions of cube?

[tsalmark]

well-broken-in, lubricated and stiff all in the same sentence, but no sex, only on slashdot.

Re:Conditions of cube?

[Mr. Underbridge]

You forgot "bigger and harder" in the title. This whole thread is just one big Freudian slit.

Re:

[L4t3r4lu5]

I semen what you did there.

Re:

[Hognoxious]

I don't get it.

Re:

[MyLongNickName]

I don't get it.

This is slashdot. You will fit in quite well here.

Re:

[tlhIngan]

Does it have to be one of those well-broken-in, lubricated cubes that easily spin, or does it work with a stiff cube just out of the packaging? I'd bet it would not.

For times well under a minute, at least a lubricated one. Even speed cubers use graphite or something. And those one-handled cubers (the ones who solve two cubes simultaneously) require it - I don't think one hand has enough agility and strength to do the twists of a brand new cube.

Mindstorm is cool and all

[tsalmark]

But that little "and a laptop" covers a whole bunch of the needed magic.

Re:Mindstorm is cool and all

[Volante3192]

Yeah, if only someone could build a robot capable of doing something without having to be programmed, that'd be AWESOME!

Re:

[tsalmark]

not so much that as: mindstorm was not the main part of the solution.

Re:

[escherblacksmith]

I have to agree with the former. the mindstorm stuff is cool, but if I am reading and watching it right, the 'solution' is handled at the software level. It would be the same as the computer telling you what moves to process to solve one. Kinda like having those old chess games that would tell you where to move its pieces as you were playing it. The laptop is the solution/brain, the mindstorms are a very slick, but inherently non-intelligent interface.

Re:

[Volante3192]

That applies to all robots though, like this:

http://hardware.slashdot.org/article.pl?sid=10/04/09/1325235 [slashdot.org]
The carving side, essentially, is dumb and just doing what HyperMill tells it to do, but without the carving side the software is useless.

To write off mindstorms simply because they require "a computer" pretty much writes off every other robot as well because they need a computer to do the job as well.

I guess I'm wondering how you would view a Willow Garage PR2 [singularityhub.com] solving the rubik's cube. Is that more 'le

Re:

[wvmarle]

And of course the actual thinking that has gone into it from both the builder(s) and the programmer(s) to make it actually be able to work. That also doesn't come from Lego.

The "only" magic that comes out of the Lego factory is of course in the form of accurate actuators, step motors, position sensors, and whatnot built into that hardware.

Re:

[n17ikh]

The same guy has built a cube solver powered by an NXT and a nokia phone [youtube.com] doing the processing, and could easily do the same with just an NXT (two, maybe?) since there is plenty of processing power there and it can even do the image recognition. People have already done this, in fact.

Re:

[cmacb]

I don't think it really counts until they can also build the computer that calculates the solution out of legos.

Re:

[edittard]

I don't think it really counts until they can also build the computer that calculates the solution out of legos.

What language will they write the softwares' codes in?

Re:

[BattleApple]

assembly language of course

Re:

[edittard]

You fail comprehensions.

Re:

[BattleApple]

Possibly.. and you fail lame humor detection.

Re:

[omglolbah]

Yep, the NXT is just acting as an IO card at this point... which you can do with any hardware really...

It takes away most of the "boring" task of doing the physical construction and circuitry for the robot. That is really all the NXT is doing...

(boring in quotes as I actually enjoy that part quite a lot..)

Comment removed

[account_deleted]

Comment removed based on user account deletion

Here's a much faster lego cube solving robot

[Rick Bentley]

sub 20 seconds, reliably: http://www.youtube.com/watch?v=eaRcWB3jwMo [youtube.com]

Re:

[Devrdander]

Thats already in the summary above, and it only does the standard cube, this one can do larger cubes.

The LAPTOP solves the cube...

[Trip6]

...the robot only actuates the solution.

Re:

[Anonymous Coward]

you must be great at parties

Re:The LAPTOP solves the cube...

[omglolbah]

I go to parties where such comments are the norm :-p

You might be surprised at the level of geekiness in some circles and how well they can party ;)

You know you have geeky friends when one of the girls refers to her clitoris as "the win button" :-p

Re:

[steelfood]

I guess her panties are the final boss that you have to take down before you can get to it.

Re:The software solves the cube...

[JaredOfEuropa]

... the laptop just runs the software.

Or is the software part of the laptop in this setup? By the same token, I think colloquial use of the word "robot" does include the computer controlling the mechanics, probably including software as well. Just thought I'd throw that in while we were being geeky ;)

Re:

[SharpFang]

Well, that's all right except a laptop isn't a lego brick.
If the whole software was running in the small controller "brick" that controls the motors, no laptop needed, I'd agree this robot was "all LEGO".

The ALGORITHM solves the cube...

[ectoraige]

...the laptop only implements the algorithm.

Re:

[ascari]

... and the algorithm is only a window into the vast solution space of Rubik's cubes ... and the...

At the risk of being assholish: Perhaps we can stop now?

Re:

[ectoraige]

I'm good.

Why is this so impressive?

[Theovon]

It's definitely cool to look at. I'll give you that. And it's cool that someone was able to take a concept involving both AI and robotics and actually build it to completion.

But from a computer science perspective, it's really not a huge deal. First off, although I personally would probably struggle a bit with the robotics because I don't do robotics, I know people who do, and they would find this robot to be mechanically rather trivial. Then the brains behind it, the algorithms to solve rubic's cubes,

That's what...

[tbcpp]

"wrote us an email to brag about its puzzle-solving might"
That's what she said!! (http://www.youtube.com/watch?v=q-wf2pP7T0Y)

Bigger cubes?

[Von Helmet]

Big deal. Call me when it can do a 4 dimensional Rubik's cube.