Newsgroups: comp.robotics
Path: brunix!sgiblab!swrinde!cs.utexas.edu!uunet!worldbank.org!markrosol-234057.worldbank.org!mark
From: Mark Rosol <mark@MarkRosol-234057.worldbank.org>
Subject: Re: Global Positioning Question
Message-ID: <1993Oct12.214036.5934@worldbank.org>
X-Xxmessage-Id: <A8E09C347E012676@markrosol-234057.worldbank.org>
X-Xxdate: Tue, 12 Oct 93 17:39:00 GMT
Sender: news@worldbank.org
Organization: The World Bank
X-Useragent: Nuntius v1.1.1d20
References: <1993Oct7.224507.10289@news.cs.brandeis.edu> <CEKvLD.Duo.2@cs.cmu.edu>
Date: Tue, 12 Oct 1993 21:40:36 GMT
Lines: 29

In article <1993Oct7.224507.10289@news.cs.brandeis.edu> Chandra Bajpai,
topgun@cs.brandeis.edu writes:
>Is it possible to have the fixed point (or several fixed) points radiate
>a radio beacon and I could triangulate my position from them?

I have developed a triangulation technique in which there are three
unique optical beacons and a rotating receiver that uniquely identifies
the location of the beacons in a room and triangulates to calculate the
robots position. This technique proved to be accurate to within a few
square centimeters in a room approx. 100 ft. square. It does require line
of site to the beacons, but is also modifiable to support more than three
beacons to allow loss of sight. You could also integrate a relative
positioning system in with the absolute positioning capability to
compensate for temporary loss of signal.

As I mentioned, the system was based on optical beacons, and research
indicated that ultrasonics or any radio frequency transmissions were
subject to multipath (signal reflection) anomalies in enclosed work
spaces, thus eliminating the ability to determine beacon location. You
may be able to employ very low frequency transmissions to eliminate much
of the multipath problem or you might try using more robust detection
algorithms to eliminate spurious reflections.

If you're interested in a copy of the paper, send me your mailing address
by email. In any event, I'd be interested in hearing a bit more about
your work.

Mark T. Rosol
mark@MarkRosol-234057.worldbank.org
