Re: [simspark-devel] status of development

[Feng X. <hen...@ma...>]

Dear all,

As the Email below said, there is a shake bug if the joint 
reaches its max/min angle and is set positive/negative
speed.

According to team's reports(the explosion problem), I think
maybe this shake bug contributes something, so I decide to
disable the shake in hingeffector.cpp and universaleffector.cpp
by:
If in min angle, negative speed will be rejected.
If in max angle, positive speed will be rejected.

These codes will change the physic performance a tiny little.
But since we will only change the vision in the next release, 
I ask for an authorization :-)




Best Regards,
Feng Xue
2008-06-11



Sent By： Feng Xue
Sent At： 2008-05-30 22:18:45
Sent To： Joschka Boedecker
CC To： Yuan Xu; Mosalam Ebrahimi; Ben
Topic： Re: status of development

Hi all,

I have refined the physics parameters now. It seems good 
according to my test. I hope it can be a RC version.


And, I came across a small problem. To make the joint move
smoothly, I have to set dFudgeParameter to a small value orelse
there will be a jitter problem. But, if this one is too small and
the joint reaches/overruns its min/max value, it will never work(
never come back again) So I have to set the dBounce to 1 to make 
it work. And, this will make the joint shake on the max/min 
position if the agent keeps sending positive/negative joint speed. 
I set the maxforce to 10 and is much better. At least, 
teams might not use this "bug" to make some strange movements :-)


I believe soccerbot has the same problems too. Talked about the
soccerbot, here is my opinion. I think we'd better make a stable,
robust RC version because it is 44 days left now. Adjusting soccerbot
will cost lots of extra energy. Allowing different robots playing 
on the same field at the same time is very great feature, but again, 
we don't have lots of time to make the match be fairly enough to both
soccerbot and Nao.


And, here is my test report of how many robots can it sim. I only 
do the test with the intergrated monitor for our Nao has reached and 
lots of things are on board now...
Machine: 
    CPU: Intel E8200 2.66GHZ
    Mem: 2G
    Graphic Card: Nvidia 8600GT
Type 1: Single Thread:
    Type 1.1: disable inner collision: can sim 7 robots smoothly
    Type 1.2: add inner collision: 3 too near robots make game unstable...
Type 2: Multi Thread:
Type 2.1: disable inner collision: can sim 7 robots smoothly
Type 2.2: add inner collision: can sim 7 robots smoothly
I have send this report to Yuan and he will do the outside monitor test.
And the result seems not good because our target is 5 vs 5 :-(


And, another thing, with the joint limit and the max joint speed feature, 
the robot is much more like the real Nao or humanoid robot. And getup is a
very difficult action. So we'd better implement an extra command like 
'beam' if all the teams failed to realize getup action :-) But no hurry,
we need feedback from teams.





Best Regards,
Feng Xue
2008-05-30



Sent By： Joschka Boedecker
Sent At： 2008-05-30 00:20:34
Sent To： Yuan Xu; Ben; Feng Xue
CC To： Mosalam Ebrahimi
Topic： status of development

Hi all,
I'm sorry that I've not been able to follow the server development too  
closely recently :-( I know that some of you have been working on the  
restricted vision perceptor, and some TC members have been asking  
about the status of its development. Could you please give me an  
update on that, and what features were implemented? Also: how about  
testing, does it work reliably?
@Yuan: Sahar has asked whether it would be possible to make the final  
release for RC08 at the end of this week. Concerning the changes that  
affect the game play, I think we have implemented everything we  
planned for, right? We can still improve the visuals until middle of  
July since they don't affect the game... Do you think you could make  
the release this weekend or early next week?
Thanks to everybody in advance for the information, and big thanks for  
your great work with the development this year!
Best regards,
Joschka