Determining Shot Accuracy of a Robotic Pool System

Abstract

A process is described to determine the shot accuracy of an automatic robotic pool playing system. The system comprises a ceiling-mounted gantry robot, a special purpose cue end-effector, a ceiling-mounted camera, and a standard bar pool table.Two methods are compared for extracting the homography between the camera and the table plane. A challenge was the relatively large area of the table surface, which required a similarly large chessboard pattern to determine the point homography. In contrast, the Dual Conic method required only a set of orthogonal lines as a calibration target, which was more convenient to manipulate, and could be calculated from the integration of multiple images with multiple target locations. The Dual Conic method was shown experimentally to recover the homography with a similar, and sometimes greater accuracy than the Chessboard method.An experimental procedure was devised to measure the accuracy of an automatic shot using a sequence of images acquired from the overhead camera. For a set of 10 shots, the average absolute angular error in placing a shot was determined to be 0.74 0 , with a standard deviation of 0.96 0 .

Toward a Competitive Pool Playing Robot: Is Computational Intelligence Needed to Play Robotic Pool?

“…This technique is awkward to apply in this case as the pattern must be very flat (typically made of glass plate) and very accurate, which is difficult using standard printing technologies. As an alternative, we have developed a method that makes use of an invariant property of the projective space which allows us to place a simple target comprising perpendicular lines, such as a large carpenter's square, at various random locations on the table [9], [11].…”

“…Finally, some of the colors are very similar and can get confused using standard color indexing methods. For example, yellow (1,9) and orange (5,13) are sometimes confused, as are red (7,15) and pink (3,11).…”

“…A camera positioned on the ceiling was used to analyze the table state and direct the robot to place a shot. More recent attempts include the Roboshark system from Iran [5], systems from Malaysia [6] and Taiwan [7], and our own Deep Green system from Queen's University, Canada [8], [9]. The Automatic Pool Trainer from the University of Aalborg in Denmark does have a vision and CI component, although it does not involve robotic actuation [10].…”

Pool Physics Simulation by Event Prediction 2: Collisions

“…The event-based method was motivated by two requirements on the simulator: that it be numerically accurate, and that it be as efficient as possible. This simulator has been developed for use within the gaming component of the Deep Green robotic pool playing system [Long et al (2004), Lam et al (2006), Greenspan (2006)] and is distributed as the physics model behind the Computational Pool Tournament of the International Computer Olympiad [Greenspan (2005)].…”