Robotics

Spring 1999 - Winter 2000. A Robot Map-Creation Algorithm. My Artificial Intelligence (CS104, Spring 1999) class project was a self-localizing, self-mapping algorithm. It grew into an ICRA paper.

Fall 1998. Paper-Manipulating Robot. I, however, did not build a lego robot; instead, I built a desktop paper-manipulating robot. It involved very interesting mechanical design and construction, and is based on a Palm III for a brain with a custom memory-mapped PIC microcontroller performing I/O.

Fall 1998. Lego Robo-Rats. In the Fall 1998 robots class, most students (not myself) made robots with Legos and the MIT Handyboard. Like Rug Warrior, the Handyboard is based on the Motorola 6811, so I was able to help provide support with Interactive C. This link contains photos of the lego robot competetion, plus some Interactive C resources.

AAAI 1997. In the summer of 1997, Tim, Will Garner, and Greg Friedland took Serial Killer to the AAAI '97 competition in Rhode Island. My role in that project was to retrofit Serial Killer with a vacuum cleaner, and to provide on-site hardware support. I believe they took second place out of five or six competitors.

Position-Encoding Floor. In April of 1997, Keith Kotay and I came up with a scheme for encoding absolute position information into a simple pixel pattern on the floor of a room, read by an inexpensive robot-mounted video camera.

AAAI 1996. Over the summer after my first robotics course, Tim, David, and I rebuilt Serial Killer in preparation for a competition at AAAI '96 in Portland, Oregon. The robot was able to perform the task successfully, but only by 3:00 AM the morning of the competition. The judges disallowed us from joining the actual competition, and instead we gave a bleary-eyed demo run. That was disappointing.

Spring 1996. The first robot I built, in the Spring 1996 Robotics course, was named Serial Killer, and based on a Newton Labs Rug Warrior kit. The other members of my team were Tim Lifshits, David Villarama, and the other David.