PITTSBURGH—A socially skilled talking robot named Grace, built by a team of researchers from Carnegie Mellon University, the Naval Research Laboratory (NRL), Metrica, Inc., Northwestern University and Swarthmore College, will participate in the American Association for Artificial Intelligence annual Robot Challenge Wednesday, July 31, in Edmonton, Alberta, Canada.
The event is part of the 18th National Conference on Artificial Intelligence (AAAI-2002) that started on July 28 and lasts through Aug. 1. (See www.aaai.org). The robot challenge on July 31 is scheduled from 1-3 p.m., with Grace beginning her mission at 2 p.m. MDT.
Grace, which stands for Graduate Robot Attending ConferencE, is a six-foot-tall, drum-shaped robot whose head is a computer monitor with a highly expressive, digitally animated female face mounted on a panning platform. She is a B21R mobile robot built by RWI, a division of iRobot, (Boston, Mass.) and contains a large array of sensors, including a microphone, touch, infrared and sonar sensors, a scanning laser range finder, and a single color camera with pan-tilt zoom capability. She can speak using a speech synthesizer and understands responses via her microphone and speech recognition software.
Grace will be dropped off at the front of the Shaw Conference Center in Edmonton. Because she has no arms, Grace will rely on the kindness of strangers to help her enter the building, where she must find her way to the registration booth, register for the AAAI meeting, and find her way to a conference room where she will give a talk about herself that includes a PowerPoint presentation.
Grace will mostly be competing against herself. A commercial device from iRobot will take part in the challenge using assisted teleoperation, and a robot developed by a team from the Massachusetts Institute of Technology will focus on the part of the challenge that deals with mapping and localization.
Carnegie Mellon's Reid Simmons, a principal research scientist in the Robotics Institute, says he'll be happy if Grace can complete even 50 percent of the tasks ahead of her. He and the other team members have spent more than a year integrating the diverse technologies necessary for Grace to perform her mission.
It's all part of a 10-year challenge to raise the bar for intelligent robot participation in a natural human environment, to stimulate robotics research groups to address these challenges and to use robotic demos to educate the public about the excitement and difficulties of intelligent robotics research.
Carnegie Mellon is the overall hardware and software integration lead on the Grace project. NRL is in charge of speech recognition and parsing, multi-modal input for speech and gesture and human-robot interaction. Metrica's part is vision-based gesture recognition, people tracking and following. Northwestern's contribution involves enabling the robot to give a talk about itself, and Swarthmore's work involves enabling Grace to read signs, read nametags and to recognize individuals.
See the Grace Web site at http://palantir.swarthmore.edu/GRACE 
For more information see:
Challenge Team 2002 (http://www-2.cs.cmu.edu/~reids/challenge/)
AAAI Mobile Robot Competition and Challenge: (http://www.cs.uml.edu/aaairobot/)
National Conference on Artificial Intelligence:(http://aaai.org/Conferences/National/2002/aaai02.html)
Byron Spice | 412-268-9068 | bspice [atsymbol] cs.cmu.edu