[Basel A]

Originally Posted by Tristan Lall

G23-1 mentions verticalness in relation to the robot. Barring the degenerate case where the entire robot is always in a fixed orientation with respect to the ground, that's proof that we can't use the ground as the sole co-ordinate reference.

In terms of using bumper edges as the reference, there are similar issues. The bottom edge of a legal bumper can (in theory) be between 2.0 in and 5.5 in from the ground, meaning significant angles are possible.

Edit: I re-read your post, and may have misunderstood it the first time. Are you suggesting that we construct the robot-relative co-ordinate system based on something like the the starting bumper orientation and the ground normal? (Hopefully the floor protector doesn't figure into this.) Then, because the bumpers can't articulate, we can use them to observe the orientation of the robot-relative co-ordinates during the climb? That has interesting implications (like when bumpers fall off), but could be feasible. However, I don't think the rules support this interpretation to the exclusion of others.

I did not spend as much time explaining my post as I could or should have, but you got the idea in your reread. I'm certainly not suggesting this is what's reflected in the current manual, but I do think this is would give results equivalent to both "common sense" and the likely interpretations of most referees. I'd challenge you or anyone else to suggest a potential rule that would be feasible and would contradict the the bumper-based model.

Note: Haven't figured out the rule that says bumpers must be "vertical." R24-E comes close, but doesn't quite do it.