There will also be another research area which some students
are invited to specialize in and that is simulated robot soccer.
The intention is to develop a soccer team which can
enter next year's Robocup in the simulator league.
The Robocup web page is at
http://www.csl.sony.co.jp/RoboCup/New/
This could also provide a basis for a physical robot
soccer team for the year after that.
The field and players are provided by a server and our
task is to provide the players' brains. There is a well-defined
set of rules and communication protocol between
robot clients and the server. The server and example client
programs are available. There is also scope here
for some 3D graphics if anyone is interested. The
simulation Robocup is two dimensional, but it would be
interesting to look at a three dimensional system, perhaps
using stored kinematic movements rather than attempting
a truly dynamical model of a soccer player.
Another line is to simulate a real robot of the type
to be used in the physical Robocup.
What I have in mind is the strategic level
of control which would probably require multiagent
techniques and intelligent control.
My brain system model is a possible
agent architecture for this, since it is
rule-based but also organized as a hierarchical
real-time control system. However, it is not necessary to
use this approach.