Path: utzoo!attcan!utgpu!news-server.csri.toronto.edu!rutgers!apple!agate!shelby!eos!jbm
From: jbm@eos.arc.nasa.gov (Jeffrey Mulligan)
Newsgroups: sci.electronics
Subject: Re: Objects in three-space
Message-ID: <7723@eos.arc.nasa.gov>
Date: 19 Dec 90 22:57:59 GMT
References: <1990Dec19.202741.24898@sbcs.sunysb.edu>
Distribution: usa
Organization: NASA Ames Research Center, California
Lines: 53

mmoss@csws8.ic.sunysb.edu (Matthew D Moss) writes:

>(* FLAME ON! *)

>  Just kidding, but I am disappointed at this group. I
>posted a very simple question a week or so ago, and didn't
>get a SINGLE response to the question. What is it, you
>don't like me? :)

I like you but my emplyer doesn't.  Some of us are supposed
to be working instead of answering questions for random strangers.
However, technology transfer is one of NASA's missions...

>  I am starting a project that will require locating some
>object (transmitter, reflector, or something) in three-space
>by way of infrared beams or the like.

One nice technique which is used for eye movement recording
is known as the "search coil" method.  The object you
want to track (say a monkey's eye) has a small assembly
mounted (with a surgical implant) that contains three pickup coils
on mutually perpendicular axes.  The animal is seated inside
an apparatus containing three large field coils that
create a magnetic field.  Each of the three field coils
is modulated at a different [high] frequency, so the signals
from the pickup coils can be demodulated using lock-in amp.
techniques.  The device operates on the principle that
the induced current in the pickup coil is proportional
to the field strength times the cosine of the angle
between the field and the coil axis.

For eye movement recording, the goal is usually to have the signal
depend only on direction of gaze, and not upon head position.
To this end, people go to some trouble to insure that the
magnetic field is very uniform.  On the other hand, if
you want to measure position, you can set up fields with
gradients and use the strength to recover position.  It
would help to have a computer if the fields are curved
with nonlinear gradients.  You might be able to do better
by doubling the number of field modulation frequencies,
and for each axis doing a differential measurement instead
an absolute measurement.

Commercial systems sell for $5-10K, I believe, but if you have
time and access to scrap parts I imaging you could put
something together for quite a bit less.  You didn't mention
what kind of accuracy you are looking for.

-- 

	Jeff Mulligan (jbm@eos.arc.nasa.gov)
	NASA/Ames Research Ctr., Mail Stop 262-2, Moffett Field CA, 94035
	(415) 604-3745