[jimfortytwo]

Paul

Yes, weight distribution is the problem with the distributed load assumption. Of course in our 6-wheel drive robot last year we didn't drop the center wheels, so it wasn't quite as egregious an omission.

Presently my hypothesis is that there is a crush load (pressure) that distinguishes wheel surfaces that just ride along top of the carpet and those that mash down into the carpet. Before reaching the crush zone, I would imagine that our basic AP physics friction rules apply, and above the crush zone I would imagine they apply too. I'm not convinced the transition zone between one and the other will be linear like the rest, however. Thats just my present thinking, however.

One problem with the white paper is that the scale of it (smallest test patch 150 square inches) makes it hard to apply with any certainty to wheels.

The main thing I'm suggesting is that tread material choice is most important, and that you can't assume either a positive or inverse relationship between surface area and friction without empirically testing your materials. (sounds roughly familiar)