Worlds Fastest (LEGO) Rubik Cube Solving Robot ?

[vinamarora8]

woooowwwww !!!!



[RoboticSolutions]

This was Cubestormer about 3 years ago,,before I added the fourth upper grip, it was the fastest scanning cube solver at the time, I moved on to create Cubestormer II with my good friend David Gilday and set the world record on 11th November 2011, which was 5.27 seconds :)

[ArmyGuy7630]

sorry the world record is 5.35 seconds including the scan

[IAssemble]

@ArmyGuy7630 you should check who designed this robot! ;-)

[bpmagruder]

Oh, you made the CubeStormer. My bad.

[bpmagruder]

Sorry, the world record is 10 seconds.

[ridzBORED]

cool robot..

watch?v=6cKfjqwiqt0&feature=wa­tch_response

[ukstevey]

that was awesome

[waawamellon]

isnt this the ONLY rubik's-cube-solving-lego-robo­t?

[MrBunnett]

did u made that lego

[RoboticSolutions]

@MrBunnett Did I build the machine? Yes. (This is an early version which developed into "CubeStormer" the 12 second cube solver.)

[gooddogsit]

i am jealous because all robots can solve the cube faster than me

[RoboticSolutions]

@gooddogsit Thats not strictly true...my NeXT robot "CubeStormer2" hasn't solved a single cube yet...maybe if you manage one solve first....?!?!?(you have about six to eight months)

[gooddogsit]

@RoboticSolutions how do you make them solve the cube?

[gooddogsit]

@RoboticSolutions how do you make them solve the cube? i would not know where to start making the robot solve the cube

[RoboticSolutions]

@gooddogsit .It's all done with clever software, "CubeExplorer", written by Herbert Kociemba, was used for this and quite a few other machines.

[chelben9]

how do u program it i am confused on how u do that i tried and i just cant program it i am using a web cam but when i play it it just scrambles it up differently

[GameMaker8Tutorial]

nice work! I have also created one, but it's very slow. The CubeStormer is the fastest, but i'm sure you can build one thats faster than it.

[RoboticSolutions]

@GameMaker8Tutorial The First version I did took almost 4 minutes. to complete a cube, with a lot of work I managed to improve the speed dramatically, what you see here is actually the lower section of CubeStormer before I added the upper gripper almost doubling its size. Cubestormer has become faster than the current video so look out for an update soon... have you posted a video of your machine? I would love to see it :)

[Whamesl0l]

@Ymate5 nope...5.6, new record xD

[Whamesl0l]

....lego, thats solves rubiks cubes?.....now ive seen it all

[meng780]

LEGO?!!?!?

[ravinago]

nerd! :)

[sajado71]

CubeStormer does it in 10 seconds...

[Spezialedbiker]

@Ymate5 wow^^but at the vid the camera doesnt help^^

[BloodXSweatXGuitars]

what....

[ToshiroKomoto]

It could be programmed to solve a specific jumbled layout. Have a random person on the street mix it up THEN try it.

[nikkadawgg]

@ToshiroKomoto agreed. this is (as the troll would put it) fake and gay :D

[antoniot157]

Wow!!

[d1995a3]

talk about home video

[VinceD41]

Holy SHizNad

[donttuchmybass]

Can you make a tutorial on how to make one!

[chickitojuanito]

Wow!

:))

[Spezialedbiker]

the world record from peoples were 24 sec

[Marco9603]

@Spezialedbiker actually, the world record is 7.08 sec

[Whamesl0l]

@Marco9603 5.6 secs actually

[SwineBrothers]

All this computer talk is hurting my head.

[5684620]

i now!

[ben1996123]

CuBear solves the cube in 6 seconds on average, so no this isnt the fastest.

[RoboticSolutions]

True,but it isnt made of lego...is it ?

[ben1996123]

Ah ok, but since you just put the word lego in brackets, I didnt know that, so yeah this probably is the fastest.

[RoboticSolutions]

No worries , btw my new 4 eight fingered version is gonna be nearly twice as fast so WATCH THIS SPACE :)

[tornadonoob1]

30-40 seconds is how fast my buddy's robot RuBot II is.

[CCxDemi923]

thats badass

[himpllx]

To me it looks as if the robot just reveresed the mixing order someone put it up as.

[Joey35826]

that is scary

[RoboticSolutions]

In practice any cube can be solved in 20 moves or less.....

software configured to download solution as soon as 20 move solve found......

i know i can execute 20 moves in under 15 seconds, so i dont want the software to waste any more time trying to find a perfect (minimum move) solution...

I havent had a cube yet that the software cant solve in a couple of seconds with more than 20 moves!

(hope this answers your question)

[pZuggler]

I don't think 20 moves is right. Last time I checked 26 moves was god's algorithm

[RoboticSolutions]

Haven't had a combination that needs more than 20 moves yet (and ive done quite a few)!!!!

[IAssemble]

@RoboticSolutions and @pZuggler - you may be interested to know that it was recently proved by a few mathemeticians and engineers (including Herbert Kociemba, author of Cube Explorer) that all positions can indeed be solved in 20 or fewer moves - see cube20 dot org for more detail if you're interested :-)

[JDoucette]

26 is all that's been proven. It's an upper limit, not a lower limit. If you download optimal cube solvers and try random positions, you'll see that 20 is the most ever required. If there are positions that require more than 20, they are very few. None have been found so far.

[pZuggler]

Wells since there are over 43 quintillion possible permutations to the cube I don't think it is valid to say that you tried a bunch of random positions and found that 20 moves was the most required. What are the odds that you tried the `position` that requires the most, or even the odds that you tried one of the `posistions`that require above average. even if you tried a third of all possible that still wouldn`t be a good indication as you would have about a 67 percent chance to havenot tried...

[pZuggler]

the one that requires the most. Since a third of all of them would be about 14 quintillion there`s no way you would ever come close to trying that many. Infact odds are that no cube will ever be in the position that requires the most moves. So there is no way you can claim that 20 is the highest because all of the ones anyone has tried have never been more, when all th ones anyone has tried would only be a small fraction of all the possible ones.

[JDoucette]

You misunderstood me. Millions, perhaps billions, of random combinations have been tried with such programs, and no one has found one requiring more than 20. Thus, the chances that a position taking more than 20 is less than 1 in a million, perhaps 1 in a billion. Every time someone used a computer to solve the optimal solution, it has produced one in 20 moves or less -- to date. It's not a proof. It's just an indicate of chances.

[pZuggler]

Actually you misunderstood me. Let's say that and enormous 50 billion random permutations have been tried (yeah right). That is only about one billionth of the possible permutations. You can't claim that because one one billionth are one way that the other 999 999 999 billionths are the same. You see the thing here is that even trying 50 billion permutations you are only trying a billionth of the total amount, and just because a billionth of them don't go over 20 moves doesn't even come close...

[pZuggler]

To indicating anything about the other 999 999 999 billionths. You are comparing 1 to 999 999 999.

[DaNovaMagic]

Mathematics has prooven that it always requires 21 moves or less.

[hubertandclyde]

Nice, I didn't hear about this, when was it proven? Must have been recent.

[DaNovaMagic]

Actually it was in the 80's. Most people here are just too fucking stupid to research.

[hubertandclyde]

Okay well I was actually at the university of waterloo in the summer of 2007 when two guys proved 26 was the most moves required, I have the news paper article too. Looks like you are wrong.

[IAssemble]

@hubertandclyde - I think you may have mis-understood the original 26 move proof. This was a proof of the "upper bound" of the value not a proof that a position exists that actually needs 26. In issue 2668 of the New Scientist magazine in August 2008 it was reported that new "upper bound" of 23 was proved. I also believe that this has now been reduced to 22. It has also been proven that there are some postions that need at least 20 moves so the maximum is somewhere between 20 and 22 inclusive.

[hubertandclyde]

Oh look, in 2008 it was proved to be 25, there's no way they proved 21 back in the 80's...

[concretecubes]

@pZuggler its more close to 23 actually.

[PaschenBack]

What is the average number of moves to solve a cube calculated by the software that you are using? Not that long ago it was proven that any cube can be solved within 26 moves, and most configurations in 23 moves I think.

[kensim28]

Is it faster than RUBOT II?

[RoboticSolutions]

yeah,its at least twice as fast as the famous Cubinator!

[AreSkavlid]

Have you ever tried taking out a piece and putting it in turned, then giving it to the robot. what do you think would happen?

[RoboticSolutions]

LOL, would it burn itself out? Malfunction....error...error! (actually the software would pick up on an invalid cube config before it came to that..!)

[bpaintball1]

wow...tes les meh dia gratz on the rubiks dubiu solveria

[ViccieB1993]

sweet!

[RoboticSolutions]

The 15 seconds I'm aiming for is for the physical solving of the cube,from starting the first face rotation to releasing the completed cube. The software I am using gives me a solution in a fraction of a second after a 5.5 second scan with the web cam. to get the solution into the robot I first have to convert the solution ie U D2 F' R L to a series of numbers that the robot can use. This is still a work in progress but the whole operation should take less than 10 seconds.

[FrappicinoINC]

so it justs does the same algorythem over and over again

[xmlisnotaprotocol]

Wow thats great!!! how long to the cube face scans take?

[RoboticSolutions]

total scan time is 5.5 seconds

[cubeCAT13]

can it do the whole entire solve planning in under 15 seconds? Including generating the solve?

[FrappicinoINC]

did u pre solve it or did the robot accualy solve it

[RoboticSolutions]

Scrambled cube scanned into pc, running cubesolver, solution downloaded to robot as a series of moves(maximum 20),robot executes moves and re-orientates the cube as fast as it can.

[FrappicinoINC]

uh.......thank u