Robotics Reference
In-Depth Information
By the time of the 2002 robot Soccer championships in Seoul, the
FIRA event had become so popular that a record 207 teams from 25
countries took part. But a rival organising group had already inaugurated
their own RoboCup Soccer championship in Nagoya in July 1997, since
when both of these organisations have promoted a worldwide prolifera-
tion of tournaments. And not only is the competition fierce between the
dozens of different universities and other research groups throughout the
world that have entered this arena, but great rivalry also exists between
the different robot Soccer organisations. It seems that politics in sport
extends to politics in robot sport.
Robot Soccer is an extremely demanding engineering activity that
requires the modelling and integration of several human skills of differ-
ent kinds: the motion skills needed for running, jumping and kicking
the ball; the perception and calculation skills necessary for seeing the
ball and the opposing players and forecasting their lines of movement;
and the mental skills that enable the robots to plan their play, strategi-
cally, tactically, and while confusing the opposition as to their intentions.
These physical and mental achievements are made possible by employing
techniques from mechanical and electronic engineering, control technol-
ogy, navigation, vision systems, communications systems and sensors,
and combining all this engineering with the intelligence of sophisticated
game-playing strategies.
In most of the tournament categories in robot Soccer, the robots are
allowed to use a vision system, based on a camera or some other type
of sensor suspended above their own half of the field. This enables the
robots to follow the ball, to know where the opposing team's robots are
located and to monitor the direction in which they are moving. In order
to detect the positions of the robots and the ball, the camera sends images
of the scene on the playing field to the robots' command computer 60
times per second. The command software then calculates the position
and orientation of its own side's robots and those of the opposing team,
identifying the players by coloured markers on the top of each robot.
Processing this data allows the command computer to forecast the paths
and locations of the ball and the opposing players through the next few
seconds, information that is needed to enable the command software to
plan its own team's immediate strategy, decide on the path each robot in
its team should be taking, and communicate this information to each of
its robots by radio. By repeating this cycle of tasks so often, the command
software is able to make its robots play as a team, co-ordinating their
