Re: Interference (was Re: Playing Field configuration?)

[Tim Vanderhoek <vanderh@ecf.utoronto.ca>]

On Sat, Jan 29, 2000 at 09:44:01PM -0500, Joyce Poon wrote:
> 
> 	Nothing is trivial...

Wiring a driver connected from a computer pin connected to a ring of
emitters is trivial.  I'm sure you can appreciate that.  :)  If I had
the parts in front of me, the soldering would take < 1 hour.  The
amount of time spent testing varies with how determined one is to find
a way to make it work.  ;-)

You're being intimidated by the electronics.  Don't.  :-)  They're
just harmless little electrons and you outweigh them by many orders of
magnitude!  In hand to hand combat, they wouldn't have a chance...


>                             I thought more about the way that detectors
> & emitters will be mounted (on the robot and on our beacon) & decided that
> the best approach is to hold off the soldering (emitters) for now.  I will
> discuss with Cindy, in person, about what I plan to do, and get her feed

Sure, just so long as you think you can keep the breadboard in one
piece for the duration of the testing.  IMHO, breadboard has a shelf-life
approximately equal to the shelf-life of plain bread without the
board.  Good luck.  :-)


> 	Um... I think our own detectors can also be confused by
> interference.  Our own sensors will be constantly picking up widely
> varying signals.

Our conclusion in a very recent mail was that picking-up signals that
are transmitted by our own robot should not be a problem, due to the
few obstacles in the field, and due to the construction of the
bounding walls.


> 	I think it's safe to assume that most grps will have squarish
> waves just because of the fact that most grps do use 555 timer.  Even if a
> non-square wave is used, the interference should still work.  

Hmmm...  I'm thinking "square" vs "rectangular".  eg, if "*" marks
times where they're emitting, and "-" marks non-emitting times,


-----***-----------***------------***----------***-----

Is much harder to interfere effectively with than

---***---***---***---***---***---***---

I think.  It's difficult to analyze quantitatively if you assume
randomly generated interference.  Actually, it's difficult to
analyze quantitatively period.


> 	I understand that. But you also have to understand that adding
> an extra subroutine in your code could mean hrs in front of the 
> oscilloscope for me.  

Hehe.  Suffice it to say that I know what is involved in building
these circuits.  I probably don't appreciate the extra time it takes
you to, for example, learn to solder, but the asymptotic value of
time spent doing things like learning to solder is zero.

The asymptotic value of time spent learning anything is usually zero,
or close to zero, provided a reasonable AI program.  :-)


-- 
Signature withheld by request of author.