Reliable and Accruate Mechanism for Launching Ping Ping Balls?

[Kurow]

Hello, I'm an electrical engineering student at Washington State University, and I am attempting to design a beer pong (more properly known as beirut) playing robot. It won't be completely autonomous, as you will still have to tell it when to shoot, and load the balls, etc. Anyway, all of the components necessary to make it are clear to me besides the launching mechanism.

What I need is a launching mechanism that will launch the ball varying distances in a reliable manner that is not too large. I've had a few ideas for doing this: something similar to a tennis ball launcher, or perhaps using compressed air? The two wheeled ball launcher is common, but according to the youtube videos I've seen, also quite inaccurate. To put the accuracy requirement in perspective: it needs to be able to get the ball inside 0.3 inches of the target position at a range of at most 10 feet. Is that possible?

I'd like to keep the robot relatively small; hopefully it will be less than a few cubic feet. Any suggestions? Thanks!

 

[Mk]

I see compressed gas used a lot, isn't it simpler to use elastic or spring energy?

 

[Kurow]

I've seen a coil gun used to propel ping pong balls, but that seems a lot more difficult to control. Compressed air looks like the most consistent solution, but I have nearly no experience with any of these things.

 

[TheRyan]

I competed in an engineering competition in high school and made a ballista out of PVC piping, a wooden crossbar and surgical tubing for propulsion of the ping pong ball. I drilled holes in the top of the PVC pipe along its length to change distance, had slots milled on both sides for the tubing to slide through and a wooden block in the middle to follow behind the ball. This was mounted on a base with a protractor-like upright with 3 angle settings. The whole thing had a 1' by 1' footprint and was ~8" high. As for the .3 leniency, if you tweak the design enough that'll come.

My design (close to the same both years, with slight modifications) was capable of accurately getting a ball in a 5 gallon bucket within 6 to 30 feet (repetitively).