It would be way cooler if it shot tiny missiles and blasted its way through.

If the first run was allowed to learn the entire maze, then this isnt really that amazing a feat. If however the first run was just an attempt to find the end, but not learn the entire maze...then thats rediculous. Id love to see their algorithm.

*quality

Boosting this quality contribution up in the Hot Listing - declared quality by ponceleon.

Yeah, as part of the competition it is allowed a slower run to solve the maze. This is the second, speed run part where it attempts to reach the centre using the known route as fast as possible.

It's still amazing purely due to the sheer speed and accuracy of the movement.>> ^charliem:

If the first run was allowed to learn the entire maze, then this isnt really that amazing a feat. If however the first run was just an attempt to find the end, but not learn the entire maze...then thats rediculous. Id love to see their algorithm.

Here is video of the first path finding run. Not quite as impressive.

www.youtube.com/watch?v=AuGe_DFHpbc&

If the Japanese had only had the advanced robotic technology that they have today back when Godzilla destroyed Tokyo, we'd all be driving Honda's and eating Sushi today.

this is still impressive as hell i think, that thing MOVES out through that maze

It looks like there's more than one path to the center, and maybe the dry run is for trying to figure out what would be the fastest path.

Even the slow run is very impressive. There was only one time I say it checking a spot that wasn't logical to check (assuming it knew the center of the maze was the goal). And when it was backtracking it was fast! I know I couldn't make it, and I've never seen one like it. That qualifies as very impressive no matter how you cut it.

>> ^SpeveO:

Here is video of the first path finding run. Not quite as impressive.
www.youtube.com/watch?v=AuGe_DFHpbc&


Wow, I found that video more impressive, at least from a programming and AI point of view.

Those movements were extremely precise even when learning, and when it traveled through a path a second time it had already calculated a much faster method.