[robocup-small] new ball.

James Bruce bruce at andrew.cmu.edu
Mon Sep 19 04:57:07 EDT 2005


Beng Kiat Ng wrote:
> I proposed to use the hockey ball about 2 to 3 years ago. But it's wasn't
> popular as with any ideas that required teams to change their robots
> drastically. Seems that teams are willing to live with the problem of golf
> ball, so long as they can use theoir current robots.

Our lab has been very involved in the dog league and its quest to 
replace the no-longer-produced Sony orange ball.  Thus we have a ton of 
potential balls around our lab ranging from ping-pong balls to half-size 
rubber soccer balls.  We tested a couple variations of hockey balls last 
Friday and we have a couple of observations:

(1) Dribblers work incredibly well on rubber balls.  Even with the 
dribblers legal by the 20% rule for golf balls, the hockey ball stuck 
extremely well to our robot.  There have been some outrageously 
effective ball holding mechanisms in the mid-size league in the past due 
to the soft and frictive outside of a soccer ball.  We'd have to change 
a lot of rules and probably adopt some of the midsize rules to combat 
overly-effective holding.

(2) For the lighter non-filled variety of floor hockey balls, we can 
kick them all the way across the field without the ball touching the 
ground (with a regular forward kicker).  I'm not sure that's what people 
meant by slowing the game down.  Obviously it doesn't move as fast as a 
golf ball would, but it's not really more controlled.

(3) For the liquid-filled damped hockey balls, we can still drive them 
2/3 of the field in the air, and they hit the ground like a brick.  Chip 
kicking these might be a little too effective, as you need not reason 
about bouncing at all.  It flies through the air, and then just rolls. 
It's true that its a slower, but chipping is easy due to the large 
relative size of the ball... all teams already know how to build a 
kicker that's only 2cm from the ground.

And now for my opinion:
Taken together, if we improve dribbling effectiveness while slowing the 
ball, it seems to me it will discourage team play.  That is, unless we 
restrict the robots' top speed at the same time (which seems to be 
unpopular and difficult to enforce).  I do think that kicking power is 
getting a bit silly, but really the >=5m/s kicks are what made passing 
preferable to driving a ball somewhere.  If we're worried about 
ref/crowd safety, we should just make those rules more explicit and 
enforce them.

That said, I *do* like the idea of exploring other ball options, but as 
in the dog league, a new ball has to be a lot better for it to be worth 
the disruption of switching.  It'd be really nice for people to bring 
example balls to RoboCup and demo them to the other teams.  I think 
that's the best way to convince a majority of teams that a new ball is 
indeed better.

Jim Bruce
CMDragons/CMRoboDragons



More information about the robocup-small mailing list