Robotics Reference
In-Depth Information
plausible amount of force applied to the cue, and the required ball would
plop neatly into the pocket, every single time. No human player would
stand a chance. But the real world is not exactly perfect. Pool tables are
not perfectly flat, friction is not precisely uniform and the other physical
features relevant to the game all have slight variations, which means that
the calculations relating to the movement of a ball when it has been hit
by the cue, and the movements of the other balls as they in turn are hit,
all contain small errors that combine to cause some balls to miss their
intended pockets.
The first attempt to design a robot snooker player was by William
Shu-Sang in the early 1990s for his PhD at Bristol University, but it
was not until a few years later that any serious interest in this field arose
within the robotics community. Michael Greenspan and his group at
Queens University in Kingston, Canada, have developed a Pool playing
robot called Deep Green, based on a gantry system. The gantry carries
an overhead camera that can view the whole table, with a second camera
used to detect and help compensate for errors in the positioning of the
gantry and the cue. Early tests with the system were all conducted on
scratch
shots—potting a single ball into a specified corner pocket, though
without the complexity of a ball-on-ball impact. Deep Green could ac-
complish this task over certain regions of the Pool table, principally near
its centre, but the robot's accuracy was diminished nearer to the edges
of the table due to errors in the automatic positioning of the gantry. By
early 2005 the robot's average success rate, based on shots from all over
the table, was around 90 percent.
Basketball
The mission of robot Basketball is to build a two-legged robot and its
control algorithm, creating a system capable of picking up the ball, mov-
ing forwards while avoiding an obstacle (another player) and then shoot-
ing the ball into the basket. This has become a popular recreational
activity at universities, particularly in the U.S.A., and is often set as a
task for student engineers. It has also been the subject of inter-collegiate
competition, for example one of the rounds of the Robot Rivals design-
and-build contest in October 2003, when teams of student engineers
from Tulane University in New Orleans and Iowa State University were
given just one day to build their robot from scratch. At the end of the
day the two robots battled it out in a match in which the robots were
