IARC Mission 2: 1996-1997
The 2nd Mission was to fly into an area where an unknown number of toxic waste drums were purported to be located. After finding each drum and mapping its location from the air relative to global positioning satellite (GPS) system coordinates, the aerial robots were then to look down and read the labels on the drums to identify the contents of each. The contents and the drum position had to be correlated. Having accomplished these tasks, the aerial robot was then to move in close to one of the drums and extract a sample (simulated by an orange metal disk located on one of the drums). Which drum contained the disk was not know beforehand.
During the 1996 event, the MIT team was able to locate all five drums and correctly identify the contents of two of them. No attempt was made to retrieve the sample. In 1996, the number of drums was known a priori (five), and the spacing of the drums was guaranteed such that no two drums could be touching. Two types of toxic waste were identified by the drum labels: Biohazardous material, and Radioactive material. For 1997, things got harder. The number of drums was an unknown. They could be touching (so the robots had to be able to distinguish the drums from the shadows that might be cast by the drums themselves), and a third type of toxic waste was added to the list: picric acid (C6H2(NO2)3 OH), a poisonous, explosive crystalline solid. The new drum label for the explosive picric acid was made to be a "morph" of the biohazard and radioactive drum labels to increase the difficulty in distinguishing each. Finally, from a mission and scoring standpoint, obtaining the sample was no longer optional.
- Auburn University: Auburn, Alabama, United States
- Carleton University: Ottawa, Ontario, Canada
- Carnegie Mellon University: Pittsburgh, Pennsylvania, United States
- DeVry-Calgary: Calgary, Alberta, Canada
- Georgia Institute of Technology: Atlanta, Georgia, United States
- Georgia State University: Atlanta, Georgia, United States
- Massachusetts Institute of Technology: Cambridge, Massachusetts, United States
- Mississippi State University: Starksville, Mississippi, United States
- Oakland University: Rochester, Michigan, United States
- Orange Coast College: Costa Mesa, California, United States
- Purdue University: West Lafayette, Indiana, United States
- Rose-Hulman Institute of Technology: Terre Haute, Indiana, United States
- Simon Fraser University: Burnaby, British Columbia, Canada
- Southern College of Technology: Marietta, Georgia, United States
- Stanford University: Stanford, California, United States
- Swiss Federal Institute of Technology (ETH): Zürich, Switzerland
- Texas Tech University: Lubbock, Texas, United States
- University of Arizona: Tucson, Arizona, United States
- University of California, San Diego: La Jolla, California, United States
- University of Central Florida: Orlando, Florida, United States
- University of Southern California: Los Angeles, California, United States
- University of Tennessee: Knoxville, Tennessee, United States
- University of Texas at Arlington: Arlington, Texas, United States
By far, the best performance was given by the Carnegie Mellon University team which based their design around a Yamaha R50 helicopter used by the Japanese for crop dusting, among other things. The Yamaha R50 is a very large and powerful machine compared to most other entries, but it allowed the Carnegie Mellon team to carry a significant amount of equipment, fuel, and batteries onboard without taxing the air vehicle. Teams using modified "hobby" helicopters as the basis for their designs are limited to payloads of about 20 lbs at most.
During the hour that Carnegie Mellon had to perform the mission, several attempts were made with the best performance demonstrating a fully autonomous takeoff, flight, and landing. During the flight all six of the drum which had been placed in the arena were correctly identified as to location and contents. But when the aerial robot tried to acquire the sample, it continued to miss its target by only centimeters. Apparently, last minute changes to the retrieval mechanism routines caused these routines to have too low a priority relative to other robotic processes (such as maintaining stability, navigation, vision, etc.), and the ensuing latency prevented the retriever from coordinating its position without offset relative to the air vehicle.
Each time the air vehicle thought it had retrieved the sample, it proceeded to return to its landing point only to find that it in fact did not have the sample. As programmed, it automatically returned to the location of the sample and tried additional retrieval attempts to no avail. When time ran out, the Carnegie Mellon aerial robot had demonstrated the most outstanding performance witnessed to date by any team entering the competition, but was unable to perform that necessary part of the mission in which a sample was to be retrieved. Faculty Sponsor, Omead Amidi stated that the entire mission had been flown several times in practice runs back in Pittsburgh, PA. Unfortunately, the two and one half centimeter offset in their robot's retrieval mechanism cost the Carnegie Mellon team $9,000 in potential prize winnings had they completed the entire mission on the day of the competition.