From: mueller@cs.unc.edu (Carl Mueller) Subject: Re: TECH: Walking in VR? Date: 19 Jun 1996 00:17:41 -0400 Message-ID: <4q7v15$111@whitney.cs.unc.edu> Organization: The University of North Carolina at Chapel Hill Kurt Wendt wrote: >James Brown wrote: >> >> I was wondering if anyone has thought of developing a VR walking >> system based on a hampster ball? [...] > > Instead of walking on the inside of a huge ball, maybe another >possibility would be to have one walk on the top of a large ball, with only >about a 2' x 2' area exposed with the rest of the ball actually located under >the floor??? The dynamics of the "over-the-ball" setup just don't work right. However, the "in-the-ball" setup works out okay for walking. With the "in-the-ball" setup, if you take a step forward, the ball automatically centers itself under your weight. This is good. With the "over-the-ball" setup, if you take a step forward, nothing special happens to recenter you. There must be a motorized setup to cause this to happen. But this becomes very complex, because you must be very careful how the force is applied. Otherwise, you will jerk the user around and he will feel this as very unnatural. As far as the "in-the-ball" setup goes, there are a variety of other problems. One obvious one is that the user must be completely enclosed: this permits no wires to go to the HMD circuitry (okay, someday we will have this all be wireless, but that's not today). Another problem happens if the user tries to run: the ball will start to move rapidly. If the user then tries to stop, the ball's mass will cause it to continue to rotate, and the user may be tossed around inside. This problem can be alleviated by making the ball very light. If one wants a non-enclosed, non-super-tall walking machine, then one must invent something similar to an escalator, but flat (like those accelerated walkways in airports) and able to operate in two dimensions (obviously the tricky part). Again, if it is to be flat, then it must be motorized, so alternatively it can be concave downward and able to move strictly by foot power. I imagine a surface made up of many small, interlinked tiles. The tiles are able to slide over another fixed surface below them. As the tiles fall off one side of the fixed surface, they are rapidly transported to the opposite side and reattached there. Of course, both the disassembly and assembly actions must be able to happen on all sides, and this makes for a very complex system. Perhaps Mr. Rubik (of cube fame) can come up with something... -Carl (mueller@cs.unc.edu)