[robocup-humanoid] Rule Discussion for 2012

[Daniel Seifert]

Hi, Dorian,

my response to some select parts of your email is below.

On Tue, Nov 15, 2011 at 18:45, Dorian Scholz <scholz at sim.tu-darmstadt.de> wrote:
>> * Goal keeper can not score a goal directly from a throw-in. Ball must
>> first
>>   touch at least one other player.
>
> A penalty should be decided upon, if the goal keeper actually throws the
> ball directly into the other teams goal.
> We think handling it in a similar way as if the ball had rolled out of the
> field over the goal line would be appropriate.
> So the ball would be placed on the half way line, but on which side has to
> be decided by the referee.

This is very close to what the TC has considered.

>> * modify game controller to count up (not down) and have the referee
>> decide
>>   the end of game (allowing to make up for game stoppages / lost time and
>>   to finish current play)
>
> We don't really see the necessity of this change, as the current GC will
> count from 10 minutes backwards and continue to count (negative) over-time.
> This allows the referee to stop the game after passing the amount of
> over-time he sees fit.

True, however there's the psychological effect that counting to 0:00
seems to mark the end. If I recall correctly, it was mentioned that
there were several games in Istanbul were the winning team started
counting down the seconds until the game was over - of course based on
the game controller, not what the referee's time was. This makes the
referees job harder and is confusing for everybody. In real-life
matches the clock also goes up, not down.

>> * increased penalty for teams that do not use the game controller
>
> We support that, but we also like to raise the question about what happens,
> if there is a wireless outage, that is not the fault of the playing teams.
> There should be a special case for that, still allowing for fluent game
> play, without the penalty being applied to the teams.

I believe an extra rule for that is not necessary. Rule 4.5.2 already
requires that robots should be able to play if the WiFi is out
(implies that they support some kind of manual start). And under
common sense, the referee would not apply any penalties if it is not
the fault of the team.

> But to allow for that it is necessary to be able to distinguish between two
> teams not supporting the GC and an actual wireless outage.
> One possibility would be to enforce the use of the GC "ping back"
> functionality, where the GC shows which robots are sending back a reply
> package.
> But for this to work the GC's network settings (subnetmask) needs to allow
> for both teams to send packets back to the GC.

Using the GC ping back definitely would be a good idea. Alternatively
it may be an option to monitor the bandwidth used (may not say much)
or if we have a display for game score as in Graz that runs on
wireless this could be used to monitor connectivity as well.

[...]

As for the technical challenges remarks, there has not been any
discussion yet on how exactly these changes are going to look like. We
appreciate your suggestions and points to consider and hope that
others will add comments of their own to allow us to define this in
the best way possible.

> Another proposition is about ensuring a high quality of the provided
> wireless network.
> In other leagues (namely Search and Rescue and the MidSize league) wireless
> network scanners are used to find sources of disturbances.
> It would be nice to have something like this in the humanoid league as well.
> Maybe in corporation with other leagues (like SPL, SmallSize) usually
> located close to our fields to split the costs/efforts and also ensure
> enforcement among their teams.

This sounds interesting. I believe the SPL has similar issues.

-- 
dseifert at fumanoids.de
Phone +49 30 838 75 125

Freie Universität Berlin
Institut für Informatik
AG Intelligente Systeme und Robotik
Arnimallee 7
14195 Berlin
Germany