The essence of our strategy involved creating a very compact, versatile, and reliable robot that would allow us to quickly finalize a physical design and leave us plenty of time to work on implementing reliable, robust software. We wanted the body of our robot to allow for a maximum set of strategic design possibilities.
In greater detail, our strategy involved attempting to reliably score the two balls nearest the starting point and several of our balls in the center of the table. With any remaining time, we would attempt to disrupt our opponent's balls and pin the balls into a wall. We had toyed with the idea of using IR tracking to attack the other robot but decided against it because of the risk of scoring our opponent's balls accidentally.
Phase I: We were able to orient with the help of IR emmiter/detector devices that determined the color of the table beneath the robot. Depending on our orientation we would turn, back up towards the gutter, knocking the first ball in, and square off with the gutter. We then drove forward following the wall nearest us, pinned the second ball, and turned quickly, "throwing" that ball into the gutter. On that run from the gutter to the second ball, we also took distance sensor measurements from sensors mounted on the sides of our robot to determine if the purple block was placed near or far from our starting point.
Phase II: We then turned and squared off against the wall behind us (facing our opponent's side), and depending on where the purple block was placed by our opponent, we would wall-follow an appropriate distance, turn towards the gutter, and plough three of our balls in. We added various angled approaches to the gutter and "shakes" to attempt to knock the balls in one at a time since they tended to get stuck in the chute if more than one was dropped at a time.
Phase III: We finally navigated our way to the opponents side of the table using some line-following and shaft-encoder-guided maneuvers, pinning several of our opponent's balls into a wall. Our servo-mounted and very stylish IR Beacon would then oscillate in celebration of victory.
Near the end of the construction phase of the competition, this entire procedure (with some exception to Phase III) was performing very reliably in all possible configurations and with many different opponents. Our last 30+ runs with our final configuration worked as intended.