BlueScream

MIT 6.270 Competition January 2001 Masters of the Universe Microsoft Team: Ya-Bing Chu ('99) Fei Chua ('94) Mike Koss ('83) Kevin Stoltz Chris Koss Web page by Mike Koss

The organizers of MIT's 6.270 invited Microsoft to participate in this year's robot competition as a corporate sponsor.  This gave us about two weeks before we had to fly to Boston, and another 5 days in the lab on campus before the day of the final competition.

We were able to demonstrate correct orientation and effective ball stealing (getting balls from the opponents side of the table and bringing them to our own side) in the lab; but we failed to do as well in the lighting conditions during the competitions.  I'd attribute this to running out of time to work the bugs out of the software - we didn't really stabilize our hardware platform and were making changes so late that we were not able to adapt and debug completely.

But we did have fun making a placebo program which we ran during contest day.  Our placebo would spin around, whack it's gate on the table, and pop wheelies.  Unfortunately (but hilariously), by placing our batteries in a spot conducive to wheelies, we also made our robot very likely to fall over on it's back; which it did in each of the demos in front of an audience.

	Friday, January 12, 2001 We received our box of robot parts.  Unfortunately, some of the (high current) batteries got loose in shipping and shorted out.  Luckily, it only charred some batteries and didn't burn up the whole box!
	Ya-Bing, Fei and Kevin at one of our evening work sessions before going to Boston.
	Kevin built a lot of the electronics and sensors used by our robot.
	Fei works on the base of our robot.
	An early version of BlueScream (at Microsoft).
	BlueScream as it appeared just before our trip to Boston.
	Chris builds our the left-hand gear train to match the design of the right.
	I put a scope on the motor speed control outputs to see what a "low speed" wave form looked like.
	Kevin and I at work at Microsoft.
	Prototype servo control for an add-on gate.  We decided that we needed to have a gate to capture balls rather than rely on being able to scoop them up in a smooth motion.
	Close up of servo mounting.
	These picture demonstrate the relative contrast of the two board colors.  The picture on the left is taken at a 45 degree angle to the board surface.  The one on the right is taken looking straight down at the surface. Note that reflection from the flash reduces the contrast between the blue and white sides of the boards noticeably; this is the reason we positioned our look-down sensor and illuminator at an angle to the table rather than have it look straight down at the table.

We programmed the robot in Interactive C.  Click here to download the final version we used in the competition.

This is the eGroups site contains an archive our email amongst ourselves and with the 6.270 coordinators.

Some more pictures from the lab.