My research centers on how everyday places are structured to facilitate common activity; e.g., buying groceries in a store, finding connecting gates in airports, or cooking in a kitchen. People plainly organize their environments, and then make use of that organization to support their activities. Likewise, artificial agents that also perform tasks in human environments (physical or virtual) can use this organization to their advantage. A good example of a structured environment is an ordinary grocery store. Stores are complex, unknown, and ever-changing places. And yet, grocery store shopping is hardly a problem for us. The reason is because we generally know how stores are structured: stores are divided into aisles, signs denote locations of items, pancake mixes can be found near maple syrups, etc. So even though we don't know the specifics of a place, we do know quite a bit to help us get along without much trouble.

Altogether, my work involves the Artificial Intelligence areas of computer vision, plan generation and execution, and robot navigation. I've integrated several capabilities---usually studied in isolation---into one whole agent. My project demonstrates how an artificial agent can make use of its knowledge of a domain in order to function adequately, optimizing its behavior as it learns the specifics of an environment. In building an agent to perform everyday tasks, I've been concentrating my efforts on these problems:

	Characterizing regularities an environment exhibits.
	Using regularities for task-relevant perception.
	Using regularities to figure out what to do now.
	Using perception to inform knowledge, and knowledge to control perception.
	Making use of past experience.
	Noticing opportunities.
	Putting it all together.

I think about these issues as they relate to everyday perception and planning. My project, Shopper, addresses these problems in performing an everyday task: grocery store shopping. You can read more about this project on the Shopper page.

I've got a number of papers written. You can read some on a page of publications, or you might want to check out the abstracts first. I don't have online versions of all my papers, so only abstracts of online papers are online.

Quicktime Movies

6 Megabytes	2 Megabytes	2 Megabytes
Shopper finds a box of Special K cereal, and then later returns to find a box of Nut & Honey nearby. You might want to read about it first though to understand what's going on.	The movie shows Shopper going into Mr. G's (the name of the store) and looking for a bottle of detergent. Totally animated. Too cool for words. Done by the great Ahmet.	This is a 3-D birds-eye view of what's happening in the previous animation. It shows Shopper's sonars sensing the surroundings. Another Ahmet production.

Artificial Intelligence Location Index	A fairly comprehensive listing of AI research groups.
Online CS Techreports	A list of technical report links to various universities.
American Association for Artificial Intelligence
International Joint Conference on Artificial Intelligence
Institute of Electrical and Electronics Engineers
Common Lisp Resources
Computer Vision Links	A listing of computer vision research groups.
CMU Face Detector Demo	Submit a picture of people and see if their faces get detected.

Computer Science Department University of Chicago 1100 East 58th Street Chicago, Illinois 60637

Last modified 16 January 1998