[robocup-small] Proposal for energy budget
James Bruce
bruce at andrew.cmu.edu
Wed Sep 14 18:09:27 EDT 2005
Sean Verret wrote:
> I very much like this proposal and would urge the TC to consider it
> perhaps not necessarily for 06 but indeed for 07. I think Raul's
> suggestion also has merit but it involves a vision system monitoring all
> the robots and monitoring their health. Not to mention, who's going to
> create the "virtual referee", my guess is that it will not prove to be
> reliable in the near future, nor is it a research topic that is
> necessarily "interesting" for most teams.
I fully agree. That's always the problem that sidelines the automatic
referee; Who's going to keep it calibrated and working on four fields?
> This suggestion by Beng will stop the arms race that has been happening
> every year for the past 5 years. Instead of coming home from Robocup
> and saying, we have to get faster, we have to build stronger kickers, we
> have to build chip kickers, we need more powerful dribblers, teams will
> come back and say, we have to pass better, we have to play as a team
> better, we have to be smarter, we have to monitor our health, etc.
> These are the "interesting" problems in my mind.
This is an unrealistic expectation in my opinion. There will always be
an arms race, but instead of being power it will be in terms of
efficiency. Teams will have capacitors (maybe even expensive
supercapacitors or ultracapacitors) to store extra battery charge so
they can temporarily outrun their opponent. Titanium and carbon fiber
will shave weight where steel and aluminum were before. Nothing much
will have changed, except competetive robots will cost even more. Just
like in professional motorcycle racing, having cc limits on motors did't
stop the hardware arms race.
Furthermore, it has negative effects on onboard computation. Allowing
seperate batteries for the computer makes the rule essentially
unenforcable; Power from "computer" batteries can easily (and even
accidentally) be routed to the motors. Limiting the total battery power
forces teams to remove substantial oboard computing power, thus making
things like local vision difficult or impossible.
That said, I do like rules that encourage more soccer-like play.
Removing the 90deg walls in 2000, outlawing side dribblers in 2004, and
limiting dribbling distance to 500mm all helped to encorage accuracy and
team play. This year we can proudly say that most of the top 8 teams
had passing which was used regularly throughout the game.
Fixing the batteries is a much more extreme change, and also makes just
about every current robot illegal. That's not to say we can't improve
things over the current situation. A few people mentioned disallowing
battery changes during the game, or lengthening the halves. Both would
go a long way toward reducing "excessive" power usage since teams would
have to lug all the batery capacity they would need for the whole game.
To make this concrete, I imagine something like the following proposal:
Every team is limited to a maximum of 7 robots and 7 battery packs for
the entire duration of the game. There would be some maximum number of
robot interchanges (3 or 4 maybe?). Each half would be 15 minutes (up
from 10). Since fewer hardware changes can be made during a game, we
can keep the overall duration of the match the same by decreasing
halftime and timeouts to 5 minutes each (down from 10).
With these rules, teams would have to ensure their robot can go for 30
minutes of play, and is robust enough that robots are not constantly
being interchanged during a game. At the same time, it doesn't enforce
a particular voltage or battery type on teams.
Jim Bruce
CMDragons/CMRoboDragons
More information about the robocup-small
mailing list