Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!qt.cs.utexas.edu!cs.utexas.edu!sdd.hp.com!ncr-sd!sagpd1!monty
From: monty@sagpd1 ()
Subject: Re: A sense of balance
Message-ID: <1993Feb9.162355.20302@sagpd1>
Reply-To: monty@sagpd1.UUCP (Monty Saine)
Organization: Scientific Atlanta, Government Products Div, San Diego, CA
References: <1993Jan31.085934.17072@adobe.com> <1993Jan31.161809.7936@sbcs.sunysb.edu> <1993Feb1.002811.22121@CSD-NewsHost.Stanford.EDU> <1993Jan31.213538.5336@memstvx1.memst.edu>
Date: Tue, 9 Feb 1993 16:23:55 GMT
Lines: 26

In article <1993Jan31.213538.5336@memstvx1.memst.edu> graesserac@memstvx1.memst.edu writes:
>In article <1993Feb1.002811.22121@CSD-NewsHost.Stanford.EDU>, jek@Xenon.Stanford.EDU (James E. Kittock) writes:
>
>There have been some interesting suggestions thrown around.  The pendulum
>ideas sound a mite mechanically messy (not to mention the alliteration :)  )
>, and everything else sounds to complicated/expensive.
>  As I was reading everyone's suggestions, it occurred to me...
>___How much accuracy could you possibly need?___
>
>If you're considering building a legged robot, you not not necessarily have
>to keep track of every little tilt the thing makes.  At most, you need to
>find the things center of gravity (just some elementary calc/physics), and
>then treat the system as a plane.  Imagine... (fade)
>

	Along the pendulum train of thought ... How about two potentiometers
    mounted at right angles to each other. Each Pot would have a weight 
    disk or arm attached. As the device was tilted the weights would 
    tend to stay down thus turning the pot(s). The accuracy would be 
    low and dampening could be adjusted by the amount of weight used. 
    An A/D converter would be used to measure the angle of tilt. Cost 
    could be trimmed in pot quality with the correspondung sacrifice
    in accuracy.


    Monty Saine
