Newsgroups: comp.robotics
Path: brunix!sgiblab!darwin.sura.net!news-feed-1.peachnet.edu!concert!sas!mozart.unx.sas.com!sasrer
From: sasrer@unx.sas.com (Rodney Radford)
Subject: Re: Bugdozer questions
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <sasrer.743643664@cinnamon>
Date: Sun, 25 Jul 1993 23:41:04 GMT
References: <22tngj$ctq@scratchy.reed.edu>
Nntp-Posting-Host: cinnamon.unx.sas.com
Organization: SAS Institute Inc.
Lines: 77

reeder@reed.edu (P. Douglas Reeder) writes:

The contest you describe is very similar to one I am working on now as
part of a rematch in our local robotics club. In our contest, the room is
approx 8'x8' and has a single 6"x6" door in the center of one wall with
a beacon placed over the door. The design of the beacon is left up to
the designer of the robot, and can be basically anything that fits in the
0.5" hole. The idea is to find and place some white blocks through the
door. The walls/floors are painted black for good contrast. There are 
actually three levels of the contest which are:

   Level 1: Find a single block and push it out the door. Contestants
            can point the robot at the block before the contest begins,
            leading to some very simple designs. (in fact, the fastest
            level 1 robot was controlled by relay logic!)
   Level 2: Find up to 4 blocks in the room and push them out the door
            in any order.
   Level 3: Push the blocks out the door in order of their number printed
            on the top of the blocks in black. Blocks should be pushed out
            in the order of 1, 2, 3, and then 4. This is the level where 
            the fun really begins!

The blocks are 1.5"x1.5", painted white, with the black numbers. The numbers
on the block use simple block characters that can be detected with as few
as 5 black/white sensors arranged in a grid.


     +-------+  +-------+  +-------+  +-------+
     |   @   |  | *   * |  | * @ * |  | * @   |
     |   @   |  | @   @ |  | @ @ @ |  | @ @   |
     |   *   |  | @   @ |  | @ * @ |  | @@*@  | 
     |   @   |  | @   @ |  | @ @ @ |  |   @   |
     |   @   |  | *   * |  | * @ * |  |   @   |
     +-------+  +-------+  +-------+  +-------+

(the '@'s represent the black paint, the '*' represent locations for the
5 white/black sensors). 

The complete rules/regulations for the contest are several pages long, but
the above is a pretty good overview of the contest. 

I just went to Toys R Us today to buy a good tracked vehicle for my design
last night, and I had the rules in front of me when I read your posting. I
thought it was similar enough that you may get some ideas for your contest.

Now on to some of your questions...

>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?

Some variation on the numbering scheme above could be used, but then you
couldn't use spherical objects (like the ping pong balls) as it would be
difficult to register them so that the numbers are easily recognized.

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

Very doubtful, unless your receiver/transmitter were locked in phase
with each other and you are prepared for some false signals. It may be
possible, but I would look for either another proximity design, or change
the beacon.

>5) Has anybody done this before?

Working on it now....   ;-}
--
---
Rodney Radford          || Computer Graphics/Imaging
sasrer@unx.sas.com      || SAS Institute, Inc.
(919) 677-8000 x7703    || Cary, NC  27513

