[robocup-small] Re: Energy budget

[James Bruce]

Hi Beng,

Beng Kiat Ng wrote:
> May I know what kind of power your robot's CPU is drawing relative to the
> whole robot power consumption? Is it much more than the other teams? I
> wonder if the CPU power consumption worry is overblown. If a CPU board
> draws 0.5A at 5 volts, a 8 cell 900mAh NiMh battery supply can  easily
> supply the CPU power  for 2 hours. Hence the main limitation is still the
> actuators' power consumption.

Well, I can tell you for at least two robots:  The CMU robots from 
2002-2003 used only about 100ma for computing, while UB Wingers' robots 
use 180ma for computing.  RoboDragons' robots use a separate 7.2V AA 
battery pack for the computer, but I don't know its consumption.  Maybe 
Tadashi can tell us.

At any rate, the consumption is small enough that it's not an issue for 
teams that I've been directly involved with; It's more an issue for 
teams that put lots of computational power onboard, such as Cornell and 
Ohio U.  My teams haven't taken that route mainly to save on cost and 
weight, but on the other hand I don't think such onboard intelligence 
should be penalized either.  It's the only way teams can experiment with 
local vision, or significant distributed AI, for example.  If local 
vision is a long-term goal or research area, then we shouldn't be 
limiting the computational power.

> Limiting batteries can introduce interesting work in SSL. A team that's not
> able to utilise the energy effectively will be penalised when it's robots'
> batteries run flat before game end. Teams can make more interesting use of
> the fresh substitution robots.

Well, what do you think of a step in that direction: Requiring a maximum 
of 7 battery packs and 7 robots per a game, with a limit of 3 
robot/battery substitutions.  I'd like to see how that works before 
taking the drastic (and harder to enforce) step of requiring a specific 
battery.  Some of the issues raised by a specific battery are: What do 
we do with teams that have CO2 powered kickers (Ohio U)?  Does that 
count toward the energy budget or not?  Can a team charge its various 
capacitors before the game?  How to we test that?

The alternative is simply to require that all the power sources be 
present before a game, and only 7 sets are allowed for the duration of 
the match (we have to distinguish power sources from robots since many 
teams do not have seven robots).

Jim Bruce
CMDragons/CMRoboDragons