Newsgroups: comp.robotics
Path: brunix!sgiblab!swrinde!elroy.jpl.nasa.gov!ames!koriel!sh.wide!wnoc-tyo-news!nkkgw!damned!hattori
From: hattori@csrd.nkk.co.jp (Mato Hattori)
Subject: Re: Robot Sumo Contest Announcement
In-Reply-To: bsmall@iat.holonet.net's message of Wed, 24 Nov 1993 19:11:58 GMT
Message-ID: <HATTORI.93Nov25120529@gaea.csrd.nkk.co.jp>
Sender: news@nkk.co.jp
Organization: NKK corp.,Tokyo,Japan
References: <HATTORI.93Nov24152844@gaea.csrd.nkk.co.jp> <CH0Fzz.Fr8@iat.holonet.net>
Date: Thu, 25 Nov 1993 03:05:29 GMT
Lines: 95


In article <CH0Fzz.Fr8@iat.holonet.net> bsmall@iat.holonet.net (Brad Smallridge) writes:

>   Thanks for the tips on Sumo Bots.  Perhaps you can tell us a little more.


You are welcome.  Sure, I will gladly try to answer any questions
about Sumo Robot.



>   One thing I was wondering about is how the robots get started. A false
>   start could give a tremendous advantage to a robot, being able to gain
>   more speed and momentum. Do the contstants start their machines by
>   flipping a switch when the judge says go?


Yes, it seems that the robot started first has an advantage.  So that
judges are severely watching the starting time in order to keep
fairness of the game.  Robots in stand alone category should have at
lease a start button and 5 sec timer.  We, operator will push the
button as soon as the judge's signal.  The robots must keep their
starting position for exactly 5 sec.  Robots may not move or even
spread their arms during that period.  Then, we have to leave at least
1 meter away from the ring during that period.  No one can be inside
the area until game ends.  If robot moved before 5 sec past since
judge's signal, the robot will lose a half point.



>   The other concern is one of friction. It would appear to me that a machine
>   that sensed when its wheels were slipping, and could adjust the power to
>   the motors accordingly, would have the advantage. The static coefficient
>   of friction is more than the dynamic coefficient of friction. 


Hmmm ... Interesting idea.  Let's try.  I have never seen robot
wrestler with precise revolution control.  All I saw was just turnning
on or off the motors.



>   You mentioned speed. Is there a way to determine the best speed? The top
>   end, I suppose, is a maximum acceleration determined by the coefficient of
>   friction times the 3KG. Hmm. Seems to be a weight factor on both sides of
>   the F=MA formula since F=Weight*Coeffictient of Friction.


It depends on how fast your CPU runs, how much inertia you have, and
how eficient your brake is.  I just intended to say that I saw many
robots had run off the ring.  I guess the reasons why they failed to
stop at the edge of the ring are probably either of poor sensors, poor
program, or too much speed.



>   The robots that ran off the ring, were they braking or using their motors
>   in reverse?


I made two Sumo robots so far.  They are braking when they want to
stop.  One uses 2 sets of MOS H bridge IC to control each motor, the
another uses 2 sets of discreet H-bridge with 4 transistors and 4
opt-isolator for each motor.  I just signaled the H bridge to be
electrically shorten turning on bottom 2 transistors and off above 2
transistors.  Reversing would be better if the robot has much inertia.
But take care of inverse voltage of inductors.  H bridge could be
easily broken with the spike voltage generated by the motors.  If you
made very-easy-maintenance robot wrestler, you may have a chance to
repair during the game.  In most case, you should consider there's no
time to exchange H bridge during the game.



Well, I'm sure building your own robot would be a great fun.  If
you're going to make one, I guess there are so much many things to do.

For example, how would you make control program for your robot?  Using
PC and cross compiler/assembler?  Yup, it would be a general way.  So,
how would you load the program onto the robot's CPU?  Burning UV-EPROM
every time you changed the program?  OK, do it if you have enough time
to develop the program.  In most case, we should have more efficient
way to download the program.  Using serial interface?  Well then, how
to load the program for serial interface?  This is an interesting
boot-strap problem, isn't it?

Sensors, interface, algorithm, ....  there're many interesting
technical points in making robots.

Enjoy your robot life!
--
Mato  Hattori                      Computer System Research Department,
email : hattori@csrd.nkk.co.jp     Electronics Research Center, NKK Corporation
voice : 81+44-322-6483 (office)    1-1 Minamiwatarida Kawasaki-ku
      : 81+44-246-9665 (home)      Kawasaki-city, 210 JAPAN
