Learn to Play like Minnesota Fats: Augmented Reality in the Pool Hall

Detecting human activities from sensors deployed in the environment is a difficult problem. Human activities are so varied that it is difficult to match sensed activities to observed actions. Instrumented homes have proved a popular platform for such sensors but experiences there show how challenging it can be. We believe that the pool hall is a more useful platform for the deployment of body-mounted sensor networks as the motions that can be captured correlate well to the activities that take place on the pool table. The problem is more narrow than detecting human activities in general. We will focus on a simple sensor, mounted on the cue, which is important as the cue is the sole actuator used to play the game. By correlating the force with which a player strikes the ball with their cue and the distance the ball travels we demonstrate that with a simple off-the-shelf sensor we can derive useful information. We conclude by describing a set of enhancements that enable a computer to “sense” the game of pool.