There were only two things that really changed (from a high level) from the beginning of our project to the end: power and catching.
We originally planned to power our 240V robot arm from tapping off two 120V circuits that were out of phase from eachother. Dr. Hurst was very justified in telling us that we could not pursue this idea. Luckily, we had just (literally 10 minutes before) found a suitable transformer that would take care of this issue for us. We implemented this transformer, and the robot worked flawlessly.
We originally planned on using the funnel that you can see to the right to dynamically catch the ball by calculating it's trajectory using some sort of dynamic method. We tried everything that we could think of, using the hardware we had available to us (which included an i7 desktop and a ZED camera), but ultimately we were not able to find a solution that could generate a position for the beach ball that the competition used (which was very unideal for Computer Vision methods) quickly enough to be useful to us. Ultimately, that meant that we just went with the method that literally every other team used - a giant cone that caught the ball.
Originally planned catching mechanism.