Re: today's (Thursday) match

Hi,

>I am using Webots under Windows-XP and it crashes after a number of
>reloads (5 to 10 times) of the Judo world and my controller.
>Currently I am testing dive attacks, and Webots never continues after
>the first hard contact.
>Webots especially seems to crash when one robot (the feet) penetrates
>(i.e. more the merely touches) the other robot.
>  
>
That sounds like a physics engine problem... Could you send me a sample 
Judoka0 program crashing Webots this way ? I would like to be able to 
reproduce this problem and eventually fix it.

>Anybody: is Webots under Linux more stable?
>  
>
I believe this problem is not platform dependant (but Iam not sure yet). 
More generally, I would say that the Linux version of Webots is a little 
bit more stable than the Windows one.

>If it is stable there, I will switch to the Linux version.
>  
>
On Linux, it is very useful if you have a nVidia graphics card to 
install the nVidia drivers from www.nvidia.com. The same applies with 
recent ATI cards where drivers can be found on www.ati.com. This allows 
for accelerated OpenGL and Webots will run much faster. On Windows, 
accelerated drivers are usually already installed.

fransklip wrote:

>While I was working on revenge against Coyote, Kimaita was surprised
>by Sonovabic winning 4 places. The latter could not win from Coyote
>either.
>
>The other big riser is Curry: went up 15 places to 6th. Very good.
>Still walking a bit slow and inefficient, but the dive attack works
>well, and the required stand-up routine equally well.
>  
>
Yes, this round was a bit surprising. It's nice to see many controller 
updates and many controllers getting closer to the top!

>Olivier:
>
>1) Very informatieve, that new match controller log!!
>With this log, you can quickly get an overview of what went on in the
>whole match round.
>  
>
Thanks.

>2) the match Sonovabic - Muhammed_ai, set 3 (the last one)
>Muhammed_ai lost this set (and the match), though Sonovabic went off
>the tatami. Reason: M. fell just before S. went off and M. did not
>recover, so M's 5 sec alarm went off first.
>You might add the rule that off-tatami causes immediate death, not
>after the 5-second recovery time-out.
>Currently, climbing back on the tatami would be allowed.
>  
>
Yes, and that would be really fun if someone could actually do it! I 
think it's all right not to change the rule about that. Climbing back to 
the tatami would be a really nice achievement.

>3) More judoka's are becoming stable in moving (walking on 2,
>sometimes 4 legs), and can recover from a fall.
>You could start thinking about a new match target/rule. Getting the
>longest distance will invite making evasive tactics.
>Perhaps give points for each time the opponent was touched (hit/kicked)?
>  
>
Hum... I don't want to hurry up to change the rules. Changing the rules 
is always a big break down in most robots' performance and I would like 
to minimize these breaks. But, it's true that more and more robots are 
now able to get up efficiently, and that's nice. The only change I plan 
to introduce is to set some limits to joint positions and to introduce 
self collision for joints and body to avoid impossible movement (like 
interpenetration).

-Olivier


------------------------ Yahoo! Groups Sponsor ---------------------~-->
Buy Ink Cartridges or Refill Kits for your HP, Epson, Canon or Lexmark
Printer at MyInks.com. Free s/h on orders $50 or more to the US & Canada.
http://www.c1tracking.com/l.asp?cid=5511
http://us.click.yahoo.com/mOAaAA/3exGAA/qnsNAA/pU_rlB/TM
---------------------------------------------------------------------~->

To unsubscribe from this group, send an email to:
[email protected]

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/