Newsgroups: comp.robotics
Path: brunix!uunet!gatech!concert!sas!mozart.unx.sas.com!sasrer
From: sasrer@unx.sas.com (Rodney Radford)
Subject: Re: It must be simple, make robot home on LED, but how?
Sender: news@unx.sas.com (Noter of Newsworthy Events)
Message-ID: <sasrer.727543023@cinnamon>
Date: Wed, 20 Jan 1993 15:17:03 GMT
Distribution: usa
References: <9161@lee.SEAS.UCLA.EDU>
Nntp-Posting-Host: cinnamon.unx.sas.com
Organization: SAS Institute Inc.
Lines: 61

yee@edison.seas.ucla.edu (John Yee) writes:

>Hey robot experts, you all must have done this a zillion times, so could
>you please relate a cheap way to make a robot home in on an LED, say from
>4 feet away inand using ich we prefer to keep lit.  

Sounds like a robot contest brewing...!

I have seen two solutions used in two seperate robot contests where the 
objective was to get the robot to within a certain distance of a visible
light source (flashing LEDs in one, non-flashing lamp in the other) that
used some sort of collimating device such that the light was only seen 
when the robot was facing directly at the light. 

********************
The first of these, just used a simple reflector from a flashlight with a 
CdS cell mounted in the focal point of the reflector (pointed back at it) 
and the robot just spun around until it found the light and drove straight 
to it. The problem with this method is that there is no easy way to make 
simple course corrections to your movements - if you loose the light, you 
just spin around until you find it again.

********************
Another method was to make a small cardboard tube with two CdS cells in the
back, and baffles in the tube (getting smaller the closer to the CdS cells).
Then you can tell which way to turn based upon how strong the light hitting
the two CdS cells are. You would just turn the robot around until you get
light on one (or both) of the CdS cells, and then you then turn the robot 
such that the light hitting the two CdS cells are equal. An ASCII sketch of
the cardboard collimator (with baffles):

+-------------------+----------------------------+-----------------------+
|                   |                            |                       |
|                   |                            |
|                   |
| O
|                       <-- light comes through here to the two
| O                             CdS cells at back of the tube
|                   |
|                   |                            |
|                   |                            |                       |
+-------------------+----------------------------+-----------------------+

In the design I saw, the tube was about 2" on a side, and over a foot long
with about 6 baffles in it. It had a very narrow detect angle and could
home in on the target very easily, and was fairly immune to ambient light. 
To help keep down reflection in the tube, be sure to paint all interior 
surfaces with a flat black paint.

********************
Of course, both of these work best when the light is at about the same
height as the detector...
Good luck.


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

