From: doug@eris.berkeley.edu (Doug Merritt)
Subject: Re: VR Sensual Feedback
Date: Tue, 5 Feb 91 17:27:46 GMT
Message-ID: <1991Feb5.172746.1195@agate.berkeley.edu>
Organization: University of California, Berkeley



In article <1991Jan16.235717.21587@agate.berkeley.edu> latta@sting.Berkeley.EDU 
(Craig R. Latta) writes:
>
>        And what about my original ideas about tactile feedback? Does anyone
>know anything about superconducting repulsion?

This is something I've been pondering for many years, and I see several
essential problems:

1) The "projected magnetic surface" cannot have a sharply defined
boundary the way that real life objects do, because the repulsion
falls off as the square of the distance. For a sharp boundary, you
need a much sharper discontinuity, where the repulsion falls from
"very high" to "very low" over a very small distance (say a millimeter).
The inverse square law instead would give a very spongey surface.
(Although see #4 below.)

2) The surface's feature resolution would be mostly two-dimensional
rather than three dimensional. At first that doesn't seem so bad;
it perhaps sounds like a spongey version of those bed-of-pins sometimes
used by artists (and hands-on museums) to make pictures. But there'd be
a further difference -- there would be little effective resistance to lateral
movements. I.e. it would be spongey side-to-side in addition to being
spongey back-to-forward.

3) Along the same lines, using the analogy of a hologram to try to
make this work better fails. There's no analogy to "collimation" and
"coherence" in these static magnetic (or electric) fields, unlike propagated
light. This means you can't have a "reference beam" to with which to create a 
reconstructed object wavefront.

4) Taking that yet one step further, all that implies the problem that
the resolution of the features in the "projected surface" falls off
with the square of the distance as well! This is probably the most
severe problem of all; it implies not just that the magnetic strength
of each "pixel" in the projecting surface must vary with the square
of the desired linear strength (an attempted fix, or one that might
help reduce lateral sponginess at the expense of forward resolution),
but further that the *absolute* "projected pixel" size will grow with
the square of the distance. (I admit to a faint hope that there may be
some lens-like mechanism that could refocus the projected "image", but I
don't currently see any way to make that work.)

5) Since we're dealing with high strength magnetic fields, there's
an additional complication of induced current flow in any metallic
objects within the field during operator/participant movement/motion.
Such problems could, for instance, burn out circuitry in the data suit.

6) Some people think that magnetic fields cause cancer or other health
problems. This may or may not be true, but it might be a complication
with your lab's insurance policy. :-)

In summary, I will not claim that there's no way to make something like
this work, but I will say that there are some extremely difficult problems
in the way.
        Doug
--
        Doug Merritt            doug@eris.berkeley.edu (ucbvax!eris!doug)
                        or      uunet.uu.net!crossck!dougm