Newsgroups: comp.robotics
Path: brunix!sgiblab!sgigate.sgi.com!olivea!charnel!charnel.net.csuchico.edu!nic-nac.CSU.net!usc!cs.utexas.edu!howland.reston.ans.net!europa.eng.gtefsd.com!MathWorks.Com!panix!zip.eecs.umich.edu!newsxfer.itd.umich.edu!gumby!yale!cs.yale.edu!ccsua.ctstateu.edu!ames_sca
From: ames_sca@ccsua.ctstateu.edu
Subject: Fire Fighting Contest info.
Message-ID: <1994Apr19.114043.1@ccsua.ctstateu.edu>
Lines: 51
Sender: news@cs.yale.edu (Usenet News)
Nntp-Posting-Host: ccsua.ctstateu.edu
Organization: Yale University, Department of Computer Science, New Haven, CT
Date: Tue, 19 Apr 1994 16:40:43 GMT

As requested I am submitting a basic outline of the construction, sensors,
and strategies for locating the candle.  First a correction on the aprox
times submitted earlier.  Mattel Hotwheels was first with 12 sec's, MIT
micro mouse with 50 sec's, third was CCSU Snufer with 58 sec's, fourth
place was a private entry called Marvin with 84 sec's.

The object of the contest was to build a robot capable of enetering a 
model four room house, locate a fire(candle), and extinguish it in the
least amount of time.  Deductions were made for not using a tether,
obstacle avoidance, returning home, etc to give more sophisticated
beasties a chance to compete with faster tether controlled machines.
A complete list of rules and spec's can be supplied upon request but
are omitted for conserving space.

The Mattel machine was untetherd but very simplistic.  It ran like a bat
out of hell and slammed into walls.  A simple 256 byte program counted
the hits and turned using a preprogrammed path.  Each time it entered a 
room it would scan using a infra-red detector and if a candle was detected
it would use a fan to blow it out.   A very fast but not practical
machine.

MIT build a very impressive, yet small, machine.  It was a wall follower 
using infra-red (walls are white, floors black, white strip across door
openings) to track the floor/wall edge.  This wasn't your typical wall
follower witha zig-zag movement.  It was extremely precise and efficient.
The mouse also used a fan to put out the candle.  We're still trying to
figure out how the hell he did it!

CCSU's machine was the most sophisticated machine but unfortunately we
based our machine on real life.  The key to the contest was speed and
simplicity.  Our machine followed a pre-programmed path, a balance beam
was located at the front, the robot would stop to adjust it's heading by
heading for a wall until two switches on the end of the beam were closed.
The beam's fulcrum is a variable resistor and is part of a tank circuit.
The robot would pivot until the circuit was balanced, thus the robot was
squared off and continue on it's way (the halls weren't wide enough to 
merit SODAR).  Snufer used infr-red to detect the candle and a combo
infra-red/opto-transister as a range finder.  Most of the cards used
were off the shelf A-Bus stuff.  We used a smart stepper controller with
acceleration\deaccel, etc.  We also installed a speech sysnth, etc that
put on a good show after the contest.

There were many impressive machines from simple lego jobs to full
featured beasties.

Further info can be obtained via E-MAIL....i've used up enough space.

                                             Scott Ames
                                             CCSU Robotics Lab


