Newsgroups: comp.robotics
Path: brunix!uunet!zaphod.mps.ohio-state.edu!mips!pacbell.com!att!news.cs.indiana.edu!umn.edu!sctc.com!smith
From: smith@sctc.com (Rick Smith)
Subject: Re: Dead reckoning
Message-ID: <1992Apr10.142455.366@sctc.com>
Organization: SCTC
References: <NIVEK.92Apr9170542@scythe.frc.ri.cmu.edu>
Distribution: comp.robotics
Date: Fri, 10 Apr 1992 14:24:55 GMT
Lines: 28

hollombe@polymath.tti.com (The Polymath) writes:

> The term usually refers to navigation by reference to known
>landmarks, ...

Well, yes, you start from a known landmark, but the "ded" for "deduced"
refers to the fact that when you need to know your location you don't
have any landmarks available. You *start* from a known landmark but
then you deduce your position based on direction and amount of motion.

My original statement was that you can't just rely on ded reckoning
("dead" reckoning). You need some kind of position sensor to close
the position control loop or the errors will accumulate.

I interpreted the original question as a search for low cost position
sensing, and I don't think that ded reconing is good enough for "low
cost" position sensing except on toys.  If you have to go to esoteric
drive systems or inertial nav, then I don't see ded reconing as a low
cost alternative to sensing external features.

You can't really ever get away from "some" ded reconing; that's
essentially what a moving device is doing as it moves between
position/feature sensor readings.

Rick.
smith@sctc.com         arden hills, minnesota


