IT takes flight in Deep Space 1

Five months after blastoff of its Deep Space 1 unmanned aircraft, NASA last week reported seeing good results from tests being conducted with cutting-edge technology.

Deep Space 1 is designed to test 12 cutting-edge technologies that NASA believes will enable the agency to build spacecraft that are smaller, cheaper to build and less dependent on human ground controllers.

The systems include an information technology-driven autonomous navigation control system and several on-board software programs that generate and execute plans for the spacecraft operations and monitoring.

The AutoNav system, which has completed about 75 per cent of its tests, “has really exceeded our expectations,” said Guy Man, chairman of the New Millennium program’s integrated product development at NASA’s Jet Propulsion Laboratory.

So far, NASA has proven the ability of AutoNav to autonomously operate the ion propulsion system and steer its course based on navigation plans stored in a computer system aboard the vehicle.

Also, the spacecraft has been able to figure out its own position by taking pictures of the sun, Earth and asteroids, measuring their position relative to background stars and comparing that data to information stored in an on-board computer.

A high-profile test of the system will be the approach and encounter with the 1992 KD asteroid in July. The navigation system is to guide the craft to within three miles of the asteroid, with a “delivery accuracy” of about 1.2 miles or better, compared to a traditional radiometric-based navigation accuracy of about 31 miles.

Man expects the AutoNav technology to yield “much better pictures than before” because the navigation camera can be pointed so precisely. The system is about 25 times more accurate than traditional methods, according to NASA.

Deep Space 1 is the first spacecraft “to tell where it is in the solar system,” said Peter Ulrich, director of the agency’s Advanced Technology and Missions Studies Division.

AutoNav technology could reduce staffing requirements for five- to 10-year missions from 30 to 10 work years, Man said.

NASA, meanwhile, has been conducting ground tests of so-called intelligent remote agent software designed to control many functions inside the spacecraft, freeing ground controllers from the work.

Remote agent software is able to plan and execute on-board activities with only general direction from the ground. It is artificial intelligence technology “pushed to the limit,” Man said. The code is intended to allow the spacecraft to respond rapidly to problems without the intervention of ground control. The remote agent software uses “model-based reasoning.”

Previously, problems on board were addressed by people on the ground who looked at data coming from the spacecraft and tried to infer what the problems were. With the intelligent code, however, information about the spacecraft is “modeled in software and flies in the spacecraft rather than on the ground in the heads of the engineers,” Man said.

The most intensive phase of testing will occur over the next few weeks, and NASA expects the software to be operating on the spacecraft by early May, Man said.

Smart software also is included in the beacon monitor system, which monitors spacecraft system health, summarizes spacecraft conditions and alerts the ground through one of four “tones,” corresponding to degrees of urgency. Beacon system validation has been 75 percent completed.

“While we’ve tested its ability to send tones, our ability to detect them on the ground and its ability to summarize spacecraft health, we haven’t yet relied on it,” said Marc Rayman, deputy mission manager for deep space at JPL.