[Guinness WR] Fastest robot to solve 3x3 cube - 2.39s!

Status
Not open for further replies.
How does it know the color of the centers? They are completely covered by the turning mechanism.
There are two possibilities:
They are unknown to the robot. So in fact it probably doesn't always solve the cube. (Judging by the video only actually nobody cared to check, that the cube is really solved at the end.)
The cube must be inserted in a fixed orientation. Then the robot has the unfair advantage (compared to cubestormer) of already knowing six of the stickers. Also in all WCA events, in which inspection time is considered as part of the solve time (blind), the cube MUST be oriented randomly. So in this case the robot cheated in my opinion.
 
How does it know the color of the centers? They are completely covered by the turning mechanism.
There are two possibilities:
They are unknown to the robot. So in fact it probably doesn't always solve the cube. (Judging by the video only actually nobody cared to check, that the cube is really solved at the end.)
The cube must be inserted in a fixed orientation. Then the robot has the unfair advantage (compared to cubestormer) of already knowing six of the stickers. Also in all WCA events, in which inspection time is considered as part of the solve time (blind), the cube MUST be oriented randomly. So in this case the robot cheated in my opinion.

Maybe there's a color sensor inside the arms that read the color of the sticker.
 
Having watched the video again and seeing, that some of the arms are labeled with a certain color, that matches the color of the (solved) adjacent cube face, I conclude almost certainly, that my second scenario must be right. The robot already knows six stickers of the cube and therefore has an unfair advantage.
 
Blindfolded rules did not apply. For this record, only articles 3 and 4 of WCA regulations are considered. It was the same way for Cubestormer. If you consider that unfair, consider this. With six arms and two cameras, all of the faces are functionally identical because they can all be turned and seen at any given time. This is unlike both other bots and human solvers. Because of this, the robot will reliably generate the same solution to the cube regardless of orientation, it will just use different arms to apply the solution. If you simply tilted your head and watched the same cube being solved from a different orientation, it would appear to be the same solve. Additionally, all scramble states are still achievable, just not with every color. The actual color representing each piece does not make a difference to the robot. In fact, colors aren't even kept track of by the computer, only the face to which they need to move toward, which is determined immediately upon scanning. Additionally, if it is the time taken to scan the six stickers (or rather two additional stickers, as that is all that is necessary to determine orientation), consider the fact that the entire scan and solution generation (more computationally expensive than scanning) is done in less than a tenth of a second. The stopwatch error alone added four tenths of a second to the time. The robot could have scanned approximately a million additional stickers in that time. Regardless, it did play by the same rules Cubestormer had.
 
Having watched the video again and seeing, that some of the arms are labeled with a certain color, that matches the color of the (solved) adjacent cube face, I conclude almost certainly, that my second scenario must be right. The robot already knows six stickers of the cube and therefore has an unfair advantage.

Just because know one has considering building it in this structure doesn't mean it's an unfair advantage. Especially if there are no rules that state that this cannot be allowed. I would call it just a simple improvement, or solution, to an easily solvable obstacle to time.
 
Status
Not open for further replies.
Back
Top