The Little Robot That Might

TJ is a robot that was built by Jon Connell at the IBM Thomas J. Watson Research Center in upstate New York (hence the robot's name - cute, huh?). It was then loaned to the MIT Aritificial Intelligence Laboratory for use by Mark Torrance as the proof of concept portion of his thesis, Natural Communication with Robots. Using the existing hardware, Mark built a rudimentary natural langauge parser and, more importantly, an implementation of "Reactive-Odometric Plans", which allowed TJ to communicate with the outside world on more human terms. After a year or so of bit-rot, it was ressurected by Sal Desiano, who made the code a little more robust, wrote a new natural language parser, and added a nifty carrying tray for TJ's Duo laptop.

The public information page for TJ was written by Mark, but is general enough that the changing details of  how TJ works don't matter. For you, dear reader, however, there are a plethora of extra details. If you want to start working on this project, here are the things you should get your hands on:

Fortunately, most of the materials in the binder are available on line (and, you don't have to retype them!):

And finally, because if all else usually fails, here are the people that have worked on the project, and an approximate e-mail address for each. Go down the list in the order given, because some people haven't worked on this project in years. But feel free to contact me (the second one on the list) even if you have a little question. I'm quite friendly.

Good luck to you. And remember, "If good lord had intended us to walk, he wouldn't have created rollerskates."

Natural Communication with Mobile Robots is a part of the Cognitive Robotics project of Professor Lynn Andrea Stein's AP group at the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology.

Page created by Salvatore Domenick Desiano,, December 7, 1996.