To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.roboticsOpen lugnet.robotics in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Robotics / 24757
24756  |  24758
Subject: 
Re: Remington: my Brickfest05 Super-Mega Sumobot
Newsgroups: 
lugnet.robotics
Date: 
Tue, 27 Dec 2005 02:07:52 GMT
Viewed: 
1447 times
  
Steve/All,

I just found this followup; I'd somehow missed it
back in October.  Many apologies, please don't read
anything into the delay...

In lugnet.robotics, Steve Hassenplug wrote:
On Tue, October 25, 2005 8:02 am, Tim Rueger wrote:

I've finally posted photos and notes on my
Brickfest05 Super-Mega Sumobot:

  http://www.brickshelf.com/cgi-bin/gallery.cgi?f=151065

Tim,

Nice robot, and you did a good job of documenting it.

I do have a question about the design.  How well did the "roller" work?

First, you have one wheel in the middle of the robot:
http://www.brickshelf.com/cgi-bin/gallery.cgi?i=1456599
http://www.brickshelf.com/cgi-bin/gallery.cgi?i=1456600

It's for watching to see if the robot is moving, right?

Do you think it's a good idea to have a free-rolling wheel that supports some of the
weight of the robot?  Wouldn't that take away from the traction the robot gets?

Yes, the roller is intended to detect (lack of)
motion.  It turns out it didn't help me much
in the competition; either I got swept off too
quickly, or I out-torqued slower opponents.

Part of the reason for the central wheel was that
my frame was flexing a lot; I hadn't considered
the net loss in traction.

And, I think there's a problem with your program.  Given this snippet:

---
#define Roller  SENSOR_3
#define RollerTimeout 50  // check rolling sensor every xx/10 seconds
...
task CheckRoller()
{
  SetPriority(RollerWatch);
  while(true) {
     // wait until the rolling timer goes off
     until(Timer(0) >= RollerTimeout);
     roll=Roller;
     if ((roll > rollwindowmin) && (roll < rollwindowmax)) {
       backupnow=1;
       SetPower(BLINK,8);
       OnFwd(BLINK); // turn on gratuitous blinker
     } else {
       backupnow=0;
       Off(BLINK);
     }

     rollwindowmin = roll-3;
     rollwindowmax = roll+3;

     ClearTimer(0);
     ClearSensor(Roller);
  }
}
---

The robot could actually be stopped for almost 15 seconds, before it will detect a
stall, not just five (the ROLLERTIMEOUT).

And, with the gearing, the robot only needs to move less than 3/10ths of an inch in
five seconds, (forward OR backward) to believe it's moving.

I didn't test this exhaustively, but when I set the
the bot against a stationary obstacle for 5 seconds,
it reverses.  It's possible the filtering/windowing code
could extend that time against an almost-non-moving
obstacle; I'll have to check that.

And I had intended to timeout only in the event of
a complete stall, so the small distance wasn't a
concern for me.  This came from my assumption that
I'd have to be pushing hard against other heavy
bots that were pushing hard (hence the 4 RC motors).

Finally, if one side of the robot is lifted just a little, the roller will come off
the ground, so it won't detect anything.

If the roller isn't moving (whether due to a
stationary obstacle or lifting), then the timer
will tell the bot something is amiss.  (Provided
the bot is still in the match at the end of the
timeout.)

Don't get me wrong.  It's clear a great deal of time went into this design.  I
suspect it was a bit over-engineered.  And, trust me, I know over-engineering.
Here's my robot from my first trip to BrickFest:

http://www.geocities.com/stevehassenplug/MyRobotPR36.html

Wow, now *that's* a Super-Mega Sumobot.  I wish I'd
seen that page before the competition, I'd have done
some things differently.

How do you think it would do against your Flatliner bot?

And yes, Remington was highly overengineered.  I
built a sumo wrestler, when I really needed a
kung fu master.

The reasons it did better at BF'02 than your robot did this year was because (A) I
had more experience at my first BF than you, and (B) the competition was much
tougher this year.

I know BrickFest was one of your first events, and it's clear you've already learned
a lot, so your sumo robots have improved, and I suspect you'll be much harder to
beat next time...

Nice work.  Keep up the good documentation.
Steve

Many thanks.  I'm looking forward to future matches.
And again, I'm very sorry I missed this reply earlier.

-Tim



Message is in Reply To:
  Re: Remington: my Brickfest05 Super-Mega Sumobot
 
(...) Tim, Nice robot, and you did a good job of documenting it. I do have a question about the design. How well did the "roller" work? First, you have one wheel in the middle of the robot: (URL) for watching to see if the robot is moving, right? Do (...) (19 years ago, 26-Oct-05, to lugnet.robotics)

2 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