NASA tests satellite software

The on-board application is designed to aim the spacecraft's cameras at interesting things on its own.

During the next several months, NASA researchers will test software designed to focus a satellite's data-gathering devices on noteworthy features and events without human prompting.

The Autonomous Sciencecraft Experiment will scrutinize Antarctica's Mt. Erebus and other volcanoes. The satellite's software analyzes images and then responds to science events, interesting features, changes relative to previous observations and cloud detection for onboard image editing. After detecting such activity, the software reprograms the camera to take additional pictures.

"This software can be used to track natural disasters that pose danger to populated areas, such as flooding and fires," said Rob Sherwood, an experiment manager at NASA's Jet Propulsion Laboratory. Future versions of the software also could track dust storms on Mars, search for ice volcanoes on Europa and track activity on Jupiter's volcanically active moon Io.

"The software is the first use of autonomy allowing the spacecraft to make decisions without waiting for commands from scientists," said Steve Chien, JPL senior technologist for the software. "It can capture short-lived science events that otherwise would have been missed." He said the software brings scientists closer to developing a thinking spacecraft.

The application, developed by JPL scientists, controls the Earth Observing-1 spacecraft. Officials at NASA's Goddard Space Flight Center manage the satellite. The software examines pictures from the Hyperion spectrometer, which is sensitive to heat from molten lava.

After taking an image of Erebus, the software detected heat from the lava lake at the summit of the volcano and reprogrammed the camera to take more pictures. Ordinarily, confirming that a remote volcano has become active takes months because without the autonomy software, scientists must take measurements at the volcano to gather the same type of information.

ASE's software is considered artificial intelligence, said Steve Chien, principal scientist of JPL's Automated Planning and Scheduling Technical Group Supervisor, Artificial Intelligence Group. It compresses several pieces, including automated planning and scheduling, to quickly detect anomalous events.

The technology could make it safer to monitor volcanoes, forest fires, and floods, according to Ashley Davies, lead scientist for the project's NMP-ST6 Autonomous Sciencecraft Experiment Asteroids, Comets and Satellites element. Autonomous spacecraft, he said, will never replace onsite investigation, but it can provide rapid alerts and precise locations for events such as an eruptions and lava flows.

"This is all useful to the people on the ground in the location of the event, to help them determine where the lava or flood water is going," Davies said.

And the software could be used for satellites that track broader phenomena, Chien said.

"We are examining many other Earth applications to track science phenomena such as dust storms, weather, ice melt, snow melt, icebergs, ocean conditions -- such as algol bloom, surf intensity -- and many others," Chien said, adding that the technology could have uses for defense and homeland security as well.

NEXT STORY: NASA uses Xerox for sharing