Performance

Our robot met the requirements, alough a couple more hours of tweaking and testing would have probably allowed us to do far better. We had two versions of code - one which simply shot free throws (we wanted to make sure we could at least fufill the bare requirements) and another which went back for some 3 pointers. In testing, the former performed quite well - we generally were able to hit 8 or 9 shots! Unfortunately, we didn't do as well with this code in the evaluation - it took 3 test runs to actually hit 3 of 9. We're unsure exactly why, although we suspect that our shooting mechanisms slipped out of place a bit. The 3 point code (whose final form hadn't been tested until the evaluation - we tested a version with different constants once, and it performed miserably...) actually did much better in the evaluation. We were able to hit 4 of 9 three point attempts (NBA caliber?). That was through a combination of mostly good last minute work (literally) and a bit of luck... And as always, the Knight Rider esque red shooting LEDs were awesome. Plus, the fact that we could exit to the buffalo command line via software rather than a reset button was helpful as well...

In retrospect, or with more time, we might have done a few things differently. Locking the shoter mechanisms into place earlier and more forcefully would have been wise, as well as paying more attention to things like how the shooting wheel was spinning and *exactly* what voltage we were using. I also would have like to add extensions to the loader for 4 more balls, which wouldn't have been hard.