
The 2003 Robot Host Contest
AAAI 2003 Mobile Robot Contest and Exhibition
August 9 - August 15, 2003, Acapulco, Mexico
Robot Host Contest Chairs: Tom Lauwers and Bruce Maxwell
Mission
The robot host competition is intended to give robotic teams (including undergraduate teams) an opportunity to demonstrate the usefulness of robots in a conference setting.
Robot Host
The objective of this contest is to create service robots that can offer information about the conference schedule, and guidance through the conference center to attendees during conference breaks.
There will be two separate tasks: a mobile information server and a robot guide. The final standings will be based on the cumulative points earned.
Registered Participants & Schedule
The list of currently registered participants is here. The schedule will be here.
Registering
If think you would like to participate (even if you are not certain), please register your interest. Participants who register early will be given the greatest consideration with regards to equipment and funding, but we will do our best to accommodate late submissions as well. Please note that there is a $250 participation fee per team. One fee covers multiple events.
Checklist:
Written summaries
To ensure that the research aspects of your exhibit are clearly understood, each entrant is strongly encouraged to submit a two page abstract summarizing your approach.
You will have an opportunity to present your work at a workshop on Friday, August 15, and write a longer paper by 1 September for publication in the workshop proceedings. Any team receiving travel support must participate in the workshop.
Competition rules and scoring
There will be two separate tasks: a mobile information server and a robot guide. Teams are not required to compete in both events to win the competition. Each competition will be scored on several metrics. The score for the mobile information server will be 70% of the total, while the score for the robot guide will be 30%.
The conference schedule and a pdf brochure are available for download now.
Most recent conference program
Format rules
Text version of conference schedule
Robots are allowed to touch attendees to move through crowds, although preference is given to robots that can accomplish this without touching. Note that hurting people by making hard or continued contact, or rolling over toes will be heavily penalized (and the robots may be banned from further competition). Consider this to be along the lines of Asimov's Rules of Robotics.
Identifying markers may be used if the markers are unobtrusive.
Finally, we want to gather data from robots engaged in both tasks. We will require that robots log sensor information and conversations during at least one run in each competition. Additionally, we plan to conduct several quantitative tests of the robot's ability to serve information/guide people. A full listing of the tests we plan to conduct will be available here.
Competition Rounds
Each competition will be divided into two events. All contestants must first pass the safety qualifier to compete in the two judged events.
Safety Qualifier
In the first round judges will evaluate if the robot is safe. If the robot "nudges" people gentleness and frequency of "nudging" will be evaluated. Judges will also check if it is impossible for the robot to run over toes. Robots that are not deemed safe will not be able to compete in any other rounds.
Mobile Information Server
The mobile information server will be stationed in the exhibit hall and will provide information about the talks and sessions at the conference. The server will greet individuals and offer them information services. The server will be tested under both high and low traffic conditions.
To be considered for scoring in this section of the competition a robot must:
- Be open to a new "conversation" with a person at any time the robot is unoccupied
- Allow information requests to be input through a human interface device (keyboard/mouse/touchscreen/microphone)
- Provide requested information visually or through speech
The environment will not be engineered for the event; the normal exhibition area will be used, although unobtrusive landmarks may be placed in the competition area. Do not assume the floor is flat. You can assume that the robot might need to pass over bare floor to carpeted areas. Furniture and other structures will be natural and may not be marked or altered for the sake of the robots. The crowd will not be limited.
Scoring for Mobile Information Server (70% of total score)
Robots will be judged in the following categories, on a 1-10 subjective scale.
- User-friendliness of the interface (35%)
- Can the robot hold a complete voice conversation?
- Is requested information displayed graphically when appropriate (maps of the conference - directions)?
- How long does it take to for the robot to understand and appropriately respond to a request?
- Robustness and Flexibility of the mobile base (10%)
- How long did the robot run during the competition? How many people did it serve?
- How often does the robot hang or crash? How often does it need to be restarted?
- Navigation (10%)
- Does the robot stay within the appropriate competition area
- Can the robot move back to near its starting position?
- Judges discretion (15%)
- Innovative designs will be rewarded
Robot Guide
The robot guide will be stationed in the main exhibition hall and will guide people towards a requested position. The guide will greet people and request if they need to be escorted anywhere. To be considered for scoring in this section, a robot must:
- Be capable of accepting a location input from a screen or through voice
- Navigate to the position requested
The environment will not be engineered for the event, although unobtrusive landmarks may be placed in the competition area. Do not assume the floor is flat. You can assume that cabling might lay on the floor, or that the robot might need to pass over bare floor to carpeted areas. Furniture and other structures will be natural and may not be marked or altered for the sake of the robots. The crowd will not be limited.
Scoring for the Robot Guide (30% of total score)
Robots will be judged in the following categories, on a 1-10 subjective scale. .
- Efficiency of the guide (15%)
- How accurately does the robot escort people?
- How easy is it for people to follow the guide? Does the guide lose people?
- Does the robot actively engage people when human traffic is low?
- Obtrusiveness of the guide system (5%)
- Are landmarks used, and how well are landmarks hidden?
- Judges Discretion (10%)
- Innovative designs will be rewarded
For More Information
... please contact us at tlauwers+@andrew.cmu.edu or maxwell+@swarthmore.edu.