Skip to Content

E.T.-Don’t Call Home

A spacecraft thinks for itself.
September 1, 1999

NASA’s fleet of robotic spacecraft has generated volumes of data about the solar system over the past 40 years. Yet, for all the insights these unmanned missions have provided, the spacecrafts are pretty dumb. A recent experiment, though, could make the next generation of planetary explorers smarter, more versatile and less expensive.

Unmanned spacecraft typically require significant handholding from mission control. A series of detailed, low-level commands are created on the ground and sent to the craft, telling it precisely what to do and when to do it. Most spacecraft also have a limited ability to deal with problems; when computers notice a problem with a key system, they shut down nonessential systems and wait for instructions from Earth.

For 35 hours over the course of one week in May, however, NASA’s Jet Propulsion Laboratory (JPL) handed over control of its Deep Space 1 spacecraft to Remote Agent, an onboard software package that uses artificial intelligence to control the spacecraft with little input from ground controllers. The Remote Agent requires only high-level commands, such as “take an image of this asteroid.” The software takes these commands and generates a timeline of events necessary to complete them. That timeline is passed on to the “Smart Executive,” which fleshes out the timeline with detailed commands, which are passed on to various subsystems. A third portion of Remote Agent, “Livingstone,” monitors the health of the spacecraft; when it detects a problem with a spacecraft system, it works to fix the problem or work around it. Only if those efforts fail does the spacecraft call home.

During the May test, Remote Agent was given a number of high-level tasks. Ground controllers also simulated problems ranging from a balky camera to a broken thruster. In each case the software handled matters on its own, despite a minor problem with the Remote Agent software that was solved within several days.

The quick fix showed Remote Agent’s versatility. “If it hadn’t been for Remote Agent’s ability to do onboard planning, we would not have been able to complete the tests so quickly,” says Pandu Nayak, deputy manager of the project.

The success is leading NASA to consider using software like Remote Agent on future missions, according to Doug Bernard, Remote Agent manager at JPL. “This technology will allow us to pursue solar system exploration missions that would have been considered too elaborate, too costly, or too dependent on teams of Earth-bound controllers,” he says.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.