So I've seen some of the omnidirectional treadmills on youtube, and they seem to work fine - but so many moving parts gives me the impression they won't be coming to the loungeroom any time soon. So I've been thinking of an alternative design, which no doubt others have considered, I just wanted to see what other musings were on the topic. It seems like it's nearly feasible, though I tried some basic tests and seems friction is far too high for it to work effectively. Say you have a spherical thick rubber bladder, reasonably elastic, maybe 10mm or more thick. You set up a short cylindrical base with a flat upper surface (the walking surface), and a kind of toroidal base - so base is hollow in middle, but still rounded on edges to minimise surface friction (final object would look a bit like a lid). This base is sealed inside the bladder and filled with lubricant, the idea being the rubber surface can be rotated easily about the base. Use omnidirectional casters to secure the base in place (so bladder can rotate while base remains steady). Then place a drive shaft wheel under the base, pushing it into the hollow region to give it greater traction. Wheel movement will make the entire surface orbit, the wheel orientation can be changed to modify the direction (using some kind of omniwheel design would be suitable as well). So does anyone think this kind of design is feasible with materials current science has available? For me it seems no matter how good your lubricant was, the friction would be too high to rotate the surface freely. Anyone else have thoughts on good way to create an omnidirectional treadmill, that doesn't have a tonne of moving parts?