This is the site of the IAP 2008 6.270 team xvi_punt.
Select a link below to find the page of your choice:
xvi_punt | Concept | Design | Construction | Competition | Your mom
Team xvi_punt is:
|
|
Kristin Uhmeyer and Matt Petersen are juniors in Course XVI. Kristin is not actually Canadian. Matt is not actually a Jawa.
The 6.270 Autonomous Robotics Design Competition takes place over IAP each January. Teams receive Lego and electronic components - controlled by the custom HappyBoard, running JoyOS - and have a month to design and build an autonomous robot to ccompete. This year, the competition involved placing balls in one of three bins, worth one, two, and four points; to total a value assigned at the beginning of each round. The robots start on random sides of the board and in random orientations. Finally, a skunk ball is available at the end of the board - if this ball winds up in one's opponent's starting square, they are docked four points.
Our first robot was the "Car of Awesomeness," which we used to test out the feasibility of gyroscopic guidance, and our motor control systems. Here it is in the 38-600 lab, receiving a program via the HappyBoard's USB interface.
At the start of each round, the robots must orient themselves in their starting squares, as they cannot receive outside input as to which side and which direction they are pointing. This is accomplished through the use of infrared LED/photostransistor pairs. These "eyes" can distinguish black and white, and thereby read the markings on the surface of the game board and send the HappyBoard information which it uses to determine the robot's position. Although the infrareds are invisible to the human eye, they are visible to the magic eye of a cell phone:
We set out to build the competition robot. We planned for a two-wheeled design, with two caster wheels and a gear ratio around 75, providing for a high speed, angry robot. The robot's guidance system was based on the gyroscope, with three sections of ball distribution - a four-ball tray to be dumped in the four-point goal; a one-ball tray to be dumped in the same, if the target score was greater than twenty, and a second one-ball tray, to be dumped in the two-point goal if the target score was 18 or 22. This was intended to produce an exact score in those two cases, and an error of at most one point otherwise.
The battery pack was mounted on the back of the robot, to provide for stability about the center of gravity. Three servos were used, each controlling one of the ball trays. After scoring int the four-point bin, the robot was to drive at high speed to the skunk ball, and crash into it, knocking it off the black line on the board and thereby rendering it un-locatable to the robots with line-following guidance systems. You can check out our software here, or, see below to find the robot loaded for battle:
Unfortunately, the competition didn't quite turn out as we planned. The robot got lost in all four rounds it played; in the first round, it was able to recover sufficiently to score, but in the others it took double losses, and was eliminated after two rounds in the finals. We didn't have time to thorougly debug the robot's sensitivity to its initial starting location before impounding; this made it pretty unpredictable, with some runs being perfect and other runs ending in getting lost. We also never tested the gyro with a long delay between the end of calibration and the beginning of the robot's run - the extra, uncompensated angle drift may have contributed to the guidance failure.
Well, not really. But we do have a modern art masterpiece: a study in perspective combining the robot, a can of Jolt, and half of the team:
