# Robot solves Rubik's Cubes in 637 milliseconds

A robot has reportedly broken the world record for solving a Rubik's Cube in the shortest time – the third occasion this year it has fallen. The builders of the Sub1 Reloaded robot claimed it took 637 milliseconds to figure out the cuboid puzzle at the Electronica trade fair in Munich. The power for motor control was supplied …

1. "Infineon claimed the test had implications for driveless cars" - now the cars can try solving Rubik's cube whilst pile driving you and your car into the side of a truck. Nice.

1. No no, the implication is that after your driver-less car piles you into the side of a truck, this robot or it's descendants will be able to examine the mess and unpick your car and the bits of your body from each other.

1. #### Unpicking the mess

So long as you and the car are in one of 43 quintillion pre-determined combinations.

1. #### Re: Unpicking the mess

And all the bits are tagged with colour coded labels, correctly oriented with respect to the robot's cameras.

2. #### Antici.......pation

“large number of unanticipated sensors”.

Somehow I think the sensors won't be unanticipated, although their data and signals quite probably will be.

Unless they're getting really advanced and taking game theory a bit too far by picking them up as they drive along as power-ups?

3. I'm more impressed with the construction quality of the cube, if it can be moved that quickly without falling apart.

1. #### I'm impressed with ...

... how quick the ink dries after being quickly sprayed on all 6 sides.

4. #### You know who you were.

Am I the only one who wants to rotate one of the corner pieces, just like we did with a lad way back when it first came out, when he claimed he could do the cube, but only when no-one was watching. Strangely his cube always felt rather loose.

1. #### Re: You know who you were.

Just to be straight, you did this to make it impossible to solve (you can also do this by flipping an edge piece)? A similar trick makes the sliding puzzle impossible to solve (switch two adjacent pieces).

1. #### Re: You know who you were.

"Just to be straight, you did this to make it impossible to solve (you can also do this by flipping an edge piece)?"

You can also pick off the colour patches and swap them around.

2. #### Re: You know who you were.

I couldn't even do the Irish O'Rubiks cube. All six sides green.

5. #### As for that time...

Is this the COMPLETE time needed to solve the cube, complete with mechanical action? Given the sheep physical distance 20 moves would take, I would call that one heck of a feat.

Which means I don't think so. Probably just the time needed to construct an efficient solution while the actual motors take it nice and easy so as not to break the cube through overexertion.

1. #### Re: As for that time...

Go to the linked BBC video. It is the actual time to solve the cube, including calculations and physical movements. Makes the cube look like it just magically changes color when played at full speed.

2. #### Re: As for that time...

Watch the video in the link......

1. #### Re: As for that time...

Ah, but they used a special low-friction cube, not the bog-standard cube humans have to solve. I'd love to know what's the fastest solving time, from first glance, using the standard cube. Also, the solution is substandard because it made 21 moves when any cube can be solved in 20 or less. So there's still room to improve.

Has anyone learned if finding the optimal solution to any Rubik's cube is considered a P or NP problem?

PS. I can't see the video due to my blockers.

1. #### Re: As for that time...

Most humans in competitions _do_ use speed cubes, which are designed to avoid unfortunate accidents like popping or corner cutting and then disassembled, lubricated, adjusted and reassembled at least twenty times during the lead-up to a competition. The first adjustment ensures that all of the cube's parts will be turning at top speed with exactly the amount of friction required while the next nineteen or so are just to give the cube's owner something to do with their hands while waiting.

2. #### Re: ....considered a P or NP problem?

P or NP?

Aaaaarrrgggghhhh.......... </runs screaming from the room>

3. #### Re: As for that time...

"Has anyone learned if finding the optimal solution to any Rubik's cube is considered a P or NP problem?"

Found my own answer, though I never knew it as "God's Algorithm". Seems an optimal solution algorithm was written in 1997, and the Cube has been exhaustively studied. But then, I have to wonder why the record breaker took 21 moves instead of 20? Was it for reasons of mechanical efficiency (easier to do certain turns in sequence than others)?

1. #### Re: As for that time...

"Was it for reasons of mechanical efficiency (easier to do certain turns in sequence than others)?"

I guess so; it may take less time to do 21 "short(er)" moves, if that is a good description for it.

4. #### Re: As for that time...

"Has anyone learned if finding the optimal solution to any Rubik's cube is considered a P or NP problem?"

I'm going to take a wild stab in dark and say you don't know what P or NP is. Because if you did you would know that solving any specific problem is very much in P. In fact in C: constant time algorithm.

1. #### Re: As for that time...

Since the robot in the video would be totally unable to solve a 4x4x4 cube, I'd say that the complexity in this case is "worse than exponential".

1. #### Re: As for that time...

"Since the robot in the video would be totally unable to solve a 4x4x4 cube, I'd say that the complexity in this case is "worse than exponential"."

Now I see what the previous commentator might have meant: the nxnxn cube, whether finding the optimal solution is in P. I very much doubt if it's in NP, and since the symmetric group on n points has order n!, I would hazard a guess it's not soluble in O(a^n) time for any a>0.

2. #### Re: As for that time...

The video that needs flash?

3. #### Re: As for that time...

Click on the link to the Beeb and you can see a video of the solution...

1. This post has been deleted by its author

4. #### Re: As for that time...

Interestingly the computation time is not included for human solvers - they get to examine the cube beforehand, and work out what they are going to do. So the 4.9 second record is simply for executing a series of pre-worked-out moves, with a few (very short) calibration stops along the route.

1. #### Re: As for that time...

"Interestingly the computation time is not included for human solvers - they get to examine the cube beforehand, and work out what they are going to do. So the 4.9 second record is simply for executing a series of pre-worked-out moves, with a few (very short) calibration stops along the route."

Which raises the possibility of a variant of the speed solve competition: from a blind start, thus taking into account on-the-spot mental solving as well as the mechanical manipulation into the time.

5. #### Re: the sheep physical distance

Is that sheep in a vaccuum or not?

6. #### “large number of unanticipated sensors”

"Oh, look here, I've got a sonar sensor! Crap, a laser rangefinder too! What am I going to do with these?"

1. #### Re: “large number of unanticipated sensors”

"Oh, look here, I've got a sonar sensor! Crap, a laser rangefinder too! What am I going to do with these?"

Attach them to sharks of course.

7. Found a video

http://www.mirror.co.uk/tech/robot-breaks-world-record-solving-9228939

1. #### World Wide Jungle

That page pulls in 3rd party content from 30 other sites and scripts from 22 other sites!

1. #### Re: World Wide Jungle

That page pulls in 3rd party content from 30 other sites and scripts from 22 other sites!

Mirror by name, mirror by nature

1. #### Re: World Wide Jungle

They probably needed to make a special robot just to fit them all onto the page in the shortest time...

2. #### Re: World Wide Jungle

Yep, and it's same old bullshit pedalled on each one... "you won't believe what happened next", "millionaires hate you knowing this secret", "this woman in X makes Y per day working at home" (stop yer tittering at the back) or "How to buy X, Y or Z for unbelievable prices" and so on. Repetition: the underpinning of brainwashing. The more "reputable" (hahaha) sites foist this bullshit on every page the more the gullible believe that any of it is remotely true and not just click bait shit.

8. ideally suited to computers programmes

9. #### Next challenge: Rubik's Magic

I still have my original one bought in 1986 from Woolies, with its box. The new silvery coloured one they brought out a couple of years back isn't very good as it just doesn't have the same tactile feel to it because it's about 2/3 the thickness of the original.

1. #### Re: Next challenge: Rubik's Magic

The Magic just isn't as fun. There's a set solution to each one, so fast solvers can do it in a second or so. At least with the Cube, the configuration can be randomized to keep things interesting.

1. #### Re: Next challenge: Rubik's Magic

Yeah I can do the switch between the three separate rings to the the interlocked rings in about 3 seconds, but the challenge here would be the dexterity of a robotic arm/hand/wotnot.

2. #### Re: Next challenge: Rubik's Magic

"At least with the Cube, the configuration can be randomized to keep things interesting."

Maybe I'm on the wrong track here, but no nitpicking intended: "... the Rubik’s cube has 43 quintillion possible combinations of coloured squares but can be completed in 20 moves." Okay, doesn't that mean that while I can pick (more or less) randomly one combination out of 43 quintillion possible combinations as the starting point, the thing itself and the process of ordering the colours is anything but random?

1. #### Re: Next challenge: Rubik's Magic

You'd be correct. Look up "parity puzzle" and you'll discover the Rubik's Cube is a type of parity puzzle (the common 15 Puzzle is another). Basically, while there are so many different configurations, those configurations are still bound by the physics of the cube that only allow certain rotations, meaning there are valid arrangements and invalid ones. The Cube has been researched quite extensively for the last couple decades.

10. You've all missed the most important point...

When do we get some of this cheesecake?

is a lie

11. #### How complex is that problem anyhow?

I mean is the problem bounded by CPU time or by how fast you can twist the cube?

1. #### Re: How complex is that problem anyhow?

It's mostly a matter of twisting the cube. It takes a tiny fraction to assess the cube and probably a touch more to determine a solution based on it. Still, the time is quick due to using a low-friction cube. Given its flimsier construction and looser tolerances, we can't expect the same results from the standard cube.

2. #### Re: How complex is that problem anyhow?

I's completely dominated by how fast you can twist the cue. You can see on the video that the movement begins almost immediately, so the visual input and computation are taking about a millisecond and then it's about 30ms per mechanical move.

It's a very nice piece of mechanical engineering that our esteemed rag appears to have mis-filed under "Artificial Intelligence".

12. This post has been deleted by its author

13. #### "he pointed out Rubrik’s cube solutions"

Yes, I hope they followed the rules.

14. #### It's an eternity!

.637 seconds....

My significant other can probably claim that I break that record, but not for solving a rubix cube....

Mine's the one with my dignity in the pocket.

15. #### Not impressed

Special cube and grabbing and rotating the centre of each face.

Whereas these use standard Rubik's cubes...

In particular...

a little different...

or, the old fashioned way

16. According to the beeb 0.637s for the FULL solution including movement http://www.bbc.co.uk/news/technology-37925028

17. #### Driverless Cars?

Driverless cars whats the connection? Oh Infineon make bits for driverless cars!

I think this might have implications for opening jam jars. Oh by the way did I mention I bought a company that makes jam jars?

El Reg can you quote me too?

18. Wow - imagine if tech like this could be transformed into something useful like machines to pick fruit, stock shelves or help elderly / disabled people to wash and cook. Tech like that might even help balance Western trade deficits if the Chinese don't develop it first...

1. Tech like that might even help balance Western trade deficits if the Chinese don't develop pirate it first...

FTFY

1. Tech like that might even help balance Western trade deficits if the Chinese don't develop pirate borrow it first...

FTFY²

1. Tsk Tsk. Has no-one let you in on the plan? Since the Chinese copy all our tech, the current strategy is to make our tech amazingly crap, diverting Chinese industry down unproductive routes whilst we carry on using old rubbish that, er, works.

Hence, Windows 10, the F35 fighter, Samsung phones, this cube solver, ...

2. The problem with picking fruit is that they are typically irregularly-shaped, positioned pretty randomly in any given tree, and a bit on the squishy side. These are problem areas where machines aren't ideal. Same with homes. They're designed for human bodies that tend to have a certain level of instinctual knowledge of how to move and so on. That's why it's been tricky to build a fully-mobile robot: some of the stuff we do comes instinctively so we don't know HOW we do it.

19. #### The World Cube Association

What a fascinating organisation name!

Nowhere near as pretentious as The Worshipful Company of Information Technologists...

20. #### "large number of unanticipated sensors"

If you have unanticipated sensors, something is very far wrong...

## POST COMMENT House rules

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