Alt-BEAM Archive
Message #12853
To: beam@sgiblab.sgi.com
From: Ben Hitchcock beh01@uow.edu.au
Date: Fri, 31 Mar 2000 12:46:50 +1000 (EST)
Subject: [alt-beam] Re: G.S.O.A.I.T.
Since the test is a measure of the 'intelligence' of the robots, how about
taking the speed of different robots into consideration.
What I mean is, find the shortest distance for the robot to achieve its
goal. Find the time taken for the robot to travel that distance, and the
intelligence of the robot is that time divided by the actual time taken by
the robot.
eg. My popper can do a good 3mm a second under a 100 W bulb. The shortest
distance to compete the course is 20 cm, so the optimum time is 66.6
seconds. My photopopper actually took 110 seconds to complete the course,
so the intelligence is 66.6 / 110 = 0.61. I think this would make it
'fairer' and would give a good measure of just how good a robot is,
regardless of the size of the solar panel or speed of the motors. If a
robot gets a score of 1 then you might want to think about patenting the
circuit!
A score of 0 means it didn't finish the course.
Just a thought,
Ben
> I'm hoping to have done some reformating and testing and more reformating of
> this series of tests within the next week. umm...i think one major thing i'll
> change is a suggestion that Bruce put up, i'm going to give categories that
> people can put their bots in, e.g. photopopper w/sensors, photopopper
> wo/sensors, beamant, solar walker, battery, no battery, etc. Anywho, i'll
> also be designing different RJP configurations as well. I plan on being as
> thorough as possible, so it might be finalized and complete by the middle of
> April. But there should be a decently upgraded update soon! THANKS A TON!
>
> -Spencer
>
> <http://www.botic.com/users/beamstop>
>
> not a robot scientist
> not a college major
> not a grad student
> not a professor
> not a very organized person
> just Spencer (isn't that impressive enough?)
>
--
Nothing is foolproof to a sufficiently talented fool.
12854 Fri, 31 Mar 2000 09:01:45 +0200 [alt-beam] Re: FRED tutorial "Thomas Pilgaard"
> I imagine that a 1381 popper would keep using that motor
> forever.. but not FRED! Occasionally, because of the difference in phase
> and period of the FLEDS, the 'wrong' motor fired, clearing the obstacle
> out from underneath and allowing FRED to run around, clearing a little
> circle of components.
Yes, this is what I have observed aswell. In my tests the FRED cirquit seems
to pop either motor somtimes, but it seems to do it consistantly - that is
kinda cyclic - according to the light it receives. For instance - motor 1
fires three times and motor 2 fires one time, then motor 1 three times and
motor 2 one time. I've had other patterns alike.
Meanwhile I've got a 1381 Popper starving for attention while I'm on the
Beam-ant and your FRED cirquit aswell. I've got a desklamp under which I'm
doing the surgery and it keeps getting in the way :)
> Better than a 1381? Probably not. But it is different, and it is
> different enough to warrant further investigation.
Indeed.
Cheers,
Thomas
>
> Ben
>
Home