To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.robotics.spyboticsOpen lugnet.robotics.spybotics in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Robotics / Spybotics / 226
225  |  227
Subject: 
Team Spybot (was: Yet another NQC spybot API submission)
Newsgroups: 
lugnet.robotics.rcx.nqc, lugnet.robotics.spybotics
Date: 
Fri, 17 Oct 2003 14:53:56 GMT
Viewed: 
718 times
  
In lugnet.robotics.rcx.nqc, Brian B. Alano wrote:

And with low-cost Spybots able to recognize and communicate with each other from
NQC, I can start playing with collective emergent behaviors (e.g. swarming). Then
the real fun begins, and the original promise of Spybots might just be possible.


I recently made a pair of Spybots that could work together for an autonomous
sumo competition.  I wanted to use the Spybot "ZONES" but found this info is not
available in NQC.

It was pretty easy to pick up Mindscript, and I was able to do just what I
wanted.

"Team Spybot" would patrol the ring, and when one would find an opponent, it
would call the other, at which point they would team up and push the opponent
out of the ring.

They weren't very strong, but they were fast.  So they were very likely to catch
the opponent from behind.

More often that not, things didn't work as planned.  One of the bots could
easily be swept off the ring.  And there were times when one found the other,
and easily swept his teammate off the ring.

I wasn't able to get any type of edge detection working, so the Spybots would
start the match by moving away from each other, and then spinning in place,
until one of them detected the opponent (using the front touch sensor).

In the competition, they could have done better.  They finished in a six-way tie
for 25th place out of 34 robots.

check it out:
http://www.24tooth.com/SumobotFive.html

BUT, when things worked, they were VERY cool to see.  First one would contact an
opponent, and call to the second.  Then, the second would home in on the IR
signal from the first, and together, they would push the opponent.

There is quite a bit of fine-tuning that I'll be able to do for next time,
mainly improving "Teammate" detection, both when pushing (is that a teammate?)
and when coming to help (which direction is my teammate)

I'm still trying to come up with other robot "projects" where I can use two or
three Spybots to work together to acomplish a task.

Steve



Message is in Reply To:
  Re: Yet another NQC spybot API submission
 
I'm hanging on every byte, John. I've tried a couple times to do what your doing with the Spybot API, and I just don't seem to have what it takes. For instance, I tried to figure out how to get the Spybot to respond to its remote control from NQC, (...) (21 years ago, 17-Oct-03, to lugnet.robotics.rcx.nqc, lugnet.robotics.spybotics)

6 Messages in This Thread:



Entire Thread on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact
    

Custom Search

©2005 LUGNET. All rights reserved. - hosted by steinbruch.info GbR