Newsgroups: comp.robotics
Path: brunix!uunet!news.cnri.reston.va.us!newsserver.jvnc.net!howland.reston.ans.net!usc!elroy.jpl.nasa.gov!decwrl!usenet.coe.montana.edu!netnews.nwnet.net!ns1.nodak.edu!plains!altenbur
From: altenbur@plains (Karl Altenburg)
Subject: Re: Bugdozer questions
Sender: usenet@ns1.nodak.edu (Usenet login)
Message-ID: <CAs8uq.t6t@ns1.nodak.edu>
Date: Mon, 26 Jul 1993 17:35:14 GMT
References: <22tngj$ctq@scratchy.reed.edu>
Nntp-Posting-Host: plains.nodak.edu
Organization: North Dakota Higher Education Computing Network
X-Newsreader: TIN [version 1.2 PL0]
Lines: 70

P. Douglas Reeder (reeder@reed.edu) wrote:
: Bugdozer Questions

: I am building a small (20 cm), robot whose goal is to find
: and return small objects to a beacon.  It is differentially
: steered with two large wheels and a smaller caster, and
: controlled by an HC11 EVB.  It will discriminate between
: objects, returning some and ignoring others; also, it will
: only return those small enough to move.  The test objects
: will be ping-pong balls and tennis balls, though any
: moveable object is fair game.  I will monitor motor current
: to detect stalls, and thus immovable objects (like walls),
: along with bump sensors and near-IR proximity detectors.

: 1) I am contemplating having the robot push the objects with
: a flexible bulldozer-blade/bump-sensor.  When pushing an
: object, the blade would be concave, to help keep the object
: in front of the robot.  When not pushing, a servo would push
: the center of the blade out, so it would be convex and less
: likely to hang up on objects and walls.  Does this sound
: feasible?  Would you suggest some alternate setup for
: pushing?  Should I just have a rigid concave blade and make
: sure the IR proximity sensors give complete coverage?

If the blade is U or V shaped you could use a break-beam sensor in the
blade to detect the presence of an appropiate sized object.  I don't know
what environment the robot is working in, but I'll assume that there
are obstacles (immovable) that would also fit in the blade.  Once something
has entered the blade, the robot could then try moving forward, and if the
object stays in the blade, and the motors don't continue to stall out for
several seconds then assume you have a 'good' object and not something like
a chair leg.  You could mount the proximity sensors on the front tips of
a U or V shaped blade.

: 2) The difference between "good" objects and "bad" objects
: must be obvious to a human, and cannot be something active
: like a beacon.  I am looking at using a phototransistor and
: visible LED to discriminate between objects, returning only
: dark ones (walls are usually light colored).  Would checking
: for some other quality be easier and more reliable?  Should
: I look for red objects, with a red filter over the sensor?

: 3) What combination and geometry of "good" object sensor,
: bump sensors, and proximity sensors should I mount on the
: blade?

: 4) Can I use the IR proximity recievers to also detect an IR
: beacon at the same 40kHz carrier without screwing up
: proximity detection?

You could modulate the beacon at some known frequency, using 40kHz as the
carrier.  For example the beacon could be modulated at 100Hz, and take
several samples to be sure it was the beacon and not external noise (such as
light bulbs, or other 40kHz IR transmissions.)

: 5) Has anybody done this before?

Using a beacon system as described? Yes, the AAAI robot building lab used the
MIT 6.270 control board and the Sharp 40kHz IR receiver in a very similar 
manner.

Bulldozing robots?  Yes, I have read and heard talks by researchers using
small robots to conduct bulldozing or fork-lifting activities, see Maja
Mataric's work for an example.

--
Karl R Altenburg				altenbur@plains.NoDak.edu
North Dakota State University, Fargo, ND  58105

All things are artificial, for nature is the art of God.  SIR THOMAS BROWNE
