June 22, 1999
NASA Ames Research Center, Moffett Field, CA
650/604-5026 or 604-9000 firstname.lastname@example.org
THREE NASA PROJECTS CHOSEN AS "DISCOVER AWARD" FINALISTS
Three NASA projects were among 27 finalists for the 10th annual Discover Awards held recently at Epcot Center, FL. Finalists were chosen in each of nine categories from an international field of thousands of entrants.
NASA's Deep Space 1 spacecraft's ion engine, managed by the Jet Propulsion Laboratory, Pasadena, CA, and called the "Revolutionary Rocket," was the winner in the Exploration category. The Lunar Prospector spacecraft that discovered water on the Moon, developed by NASA Ames Research Center, Moffett Field, CA, was a runner-up in that category, and another Ames project, experimental "smart airplane" software was a finalist in the Transportation category.
The smart plane software which can help pilots safely land aircraft that have suffered major failures was flight tested on a modified F-15 aircraft. Each sixth of a second, a damaged aircraft's computer can "relearn" to fly the aircraft using special neural network "controller" software. Without the smart software, severe problems such as partially destroyed wings, major fuselage tears or sensor failures can greatly alter how an airplane handles, and the aircraft might respond oddly or pilots' controls may not work properly.
"We were fascinated when we saw each others' inventions during the Awards events," said Dr. Chuck Jorgensen, a NASA Ames scientist who leads the smart plane software effort. "I felt very honored when I saw how many amazing inventions competed."
In the Exploration category, "To the Moon, Cheaply," the Lunar Prospector spacecraft project, was runner-up to the Revolutionary Rocket. Lunar Prospector exemplifies NASA's new way of doing business, having set new standards in cost containment and schedule for a NASA exploration mission.
Developed in less than three years at a total cost of $63 million, Lunar Prospector provided evidence that water ice exists in the permanently shadowed craters of the lunar polar regions. Prospector also yielded data that have led to the development of complete gravity maps of the Moon, maps of location and extent of key minerals and other elements, and evidence concerning tectonic and volcanic activity.
"I want to add my congratulations to all those who are working tirelessly to develop new innovations and technologies to meet the needs of people worldwide in the next century," Vice President Al Gore wrote in a letter to awards ceremony attendees. "Your work is critical to the success of our country."
Jorgensen noted that neural network software being developed in the "smart plane" project could have a bearing on other aspects of contemporary life. "Once we prove neural net software can rapidly learn to fly a crippled aircraft and help pilots land it safely, engineers will be more likely to use the intelligent neural network software in power plants, automobiles and other less-complicated systems to avoid potential disasters after equipment failures," he said.
The first flight tests of Jorgensen's Intelligent Flight Controller took place at NASA Dryden Flight Research Center, Edwards, CA, using early versions of the new software installed in a modified F-15 jet fighter. The Boeing Company's Phantom Works division, St. Louis, MO, integrated the NASA Ames neural network software into the F-15 test aircraft. Jorgensen is the principal investigator for the four-year Intelligent Flight Controller Program at Ames.
"Neural net software learns by observing 'patterns' in the real world and learning to take new actions in response to different patterns," Jorgensen said.
The software gets speed, direction and force data from sensors on the aircraft. The aircraft's computer compares the pattern of what is actually happening to the aircraft with a pattern showing how the airplane should fly. If there is a mismatch, the computer software, which contains a dozen basic aeronautical equations, or "behavior patterns" that define how airplanes fly, makes the system work with a "new pattern," if it is feasible.
"If sensor data show that a pattern is not being followed, and the airplane is turning too abruptly, the airplane's neural network can rapidly learn to assist the pilot. It does this by helping the pilot to use the stick, flaps, rudders and other control surfaces in ways that may be very unconventional, but possibly successful," said Jorgensen.
- end -
To receive Ames news releases via e-mail, send an e-mail with the word "subscribe" in the subject line to email@example.com. To unsubscribe, send an e-mail to the same address with "unsubscribe" in the subject line.