Research Links

AI Research (heavily influenced by a couple of folks at MIT, if you can imagine)

• Rodney Brooks' work, including the subsumption architecture, struck some people as revolutionary way back when Brooks was still making robot bugs. He's making an android the last I heard, but his work is still worth reading up on. In particular, read "A Robust Layered Control System for a Mobile Robot"
• Pattie Maes' work builds on Brooks' subsumption architecture, to some degree. My Bachelors thesis was written under Pattie's supervision. Amongst all she's done, the paper that I suggest reading is "How To Do the Right Thing". This paper describes an "activation net" of competency modules. The notion is that different modules inhibit or encourage other modules to activate to satisfy predecessor or exclusion relationships between activities. No, it's not a neural net.
• Bruce Blumberg is heading the Synthetic Characters group at the MIT Media Lab. He began studying under Pattie shortly after I did, and he has carried on the torch of situated artificial life intelligence. He and his group have a bunch of fun papers here.

Physics

• Advanced Character Physics by Thomas Jakobsen. This page describes an approach to physics that doesn't require solving the equations of motion for a 3-d body, and yet still yields believable results with quite acceptable speed. For what it's worth, I showed this stuff to Matt Scott, who then won a PS2 contest making a marionette that ran in the VU using this approach.

Terrain

• Binary Triangle Trees, as seen in Tread Marks, this page describes an inexpensive way for terrain to be represented so that you can do continuous level of detail (CLoD) on dynamic terrain.
• Thatcher Ulrich's Terrain System caught my attention back when it was a GamaSutra article in anticipation of SoulRide. Check out the original paper here, or look at his other stuff here.