Magazine article AI Magazine

The 2000 AAAI Mobile Robot Competition and Exhibition

Magazine article AI Magazine

The 2000 AAAI Mobile Robot Competition and Exhibition

Article excerpt

A popular attraction at the 2000 AAAI National Conference on Artificial Intelligence was the Ninth AAAI Robot Competition and Exhibition, held 30 July to 3 August 2000 in Austin, Texas. This year's event brought six contest teams and nine exhibition teams from the United States and Canada.

The Robot Contest and Exhibition brings together teams from universities and other laboratories to compete and demonstrate state-of-the-art research in robotics and AI figure 1.

[Figure 1 ILLUSTRATION OMITTED]

The contest and exhibit have several goals: (1) encourage students to enter robotics and AI fields at both the undergraduate and graduate levels, (2) increase awareness of the field, and (3) foster the sharing of research ideas and technology.

The competition and exhibition is actually made up of multiple events: several contests, a challenge event, an exhibit, and a final workshop for all participants. Descriptions of previous years events can be found in Dean and Bonasso (1993); Konolige (1994); Simmons (1995); Hinkle, Kortenkamp and Miller (1996); Kortenkamp, Nourbakhsh, and Hinkle (1997); Arkin (1998); and Meeden et al. (2000).

The Competition: Something Old, Something New

The competition this year consisted of two events: Hors d'Oeuvres, Anyone? and a new event, Urban Search and Rescue.

Hors d'Oeuvres, Anyone?

The objective of the Hors d'Oeuvres, Anyone? contest is to create service robots that can offer hors d'oeuvres to attendees at the reception. The event stresses human-robot interaction, as well as mobility, and each contestant is required to explicitly and unambiguously demonstrate interaction with the spectators.

The fourth year for this popular event, the robots are judged while they serve finger foods to attendees at the AI Festival. Unlike other contests over the years, there were no artificial walls or constraints in this event--the robots had to interact with regular conference participants, and no attempt was made to limit the number of people interacting with each robot. Robots were judged on the quality of their interactions, coverage, and ability to refill their trays (such as detecting when they needed a refill and navigating to a refill station).

First place in this event went to Swarthmore College, which received an AMIGOBOT from ActivMedia. An article by the winning team, which better describes their approach and robot, can be found in this issue of A1 Magazine. Second place was awarded to the University of Arkansas, which received a MAGELLAN robot from Real World Interface (RWI).

Urban Search and Rescue

In January 2000, a suggestion was made to introduce a new contest, Urban Search and Rescue (USAR). There has been a growing interest in this humanitarian task, a task whose complexities require much greater integration of AI techniques, such as mobility in three dimensions, the ability for the robot platform to change shape, adjustable autonomy, robot-human interaction, sensor fusion, and accurate map building and localization (Blitch, Sidki, and Durkin 2000; Blitch 1996).

Being late in the year, I was uncertain how many teams might be attracted to the contest, but I felt that even with a small turnout, this event could be used to introduce the task to the AI community. Although a few teams came with very different approaches, the event drew attention and brought forward many teams that say they will compete in 2001.

The USAR arena (figure 2) consisted of three sections, which differed in physical complexity from the easiest section, with a flat-world assumption, to the hardest area, which included small spaces, nonflat and unstable surfaces, and debris. Figures 3 to 5 show robots from the University of South Florida (USF) in the two more difficult sections of the arena.

[Figures 2-5 ILLUSTRATION OMITTED]

One of the most difficult aspects of this event was to generate metrics to be used in judging the event. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.