JPL Leads NASA’s Breakthrough in Autonomous Satellite Imaging

NASA has successfully tested an artificial intelligence system that allows satellites to autonomously decide when and where to capture images. The technology, called Dynamic Targeting, was developed by the Jet Propulsion Laboratory in Pasadena and deployed aboard CogniSAT-6, a CubeSat built by Open Cosmos.

The July flight demonstrated the system’s ability to complete a full scan-decide-capture loop in just 60 to 90 seconds, even while traveling nearly 1,300 kilometers down-track at orbital speeds. It works by tilting the satellite forward 40 to 50 degrees to scan for cloud cover before pitching back to take images only of clear scenes.

CogniSAT-6 uses a single optical sensor for both targeting and imaging, powered by an Intel Myriad X processor executing cloud detection algorithms in under five seconds. The AI software was developed at JPL over a decade of research, with funding from NASA’s Earth Science Technology Office. NASA’s press release dated July 24 confirmed the successful test and outlined the system’s operational goals.

“The idea is to make the spacecraft act more like a human: Instead of just seeing data, it’s thinking about what the data shows and how to respond,” said Steve Chien, JPL technical fellow in AI, in the NASA announcement.

Open Cosmos provided spacecraft operations and mission services under a multi-year contract. Dublin-based Ubotica supplied the AI hardware, previously tested aboard the International Space Station in 2021.

“If you can be smart about what you’re taking pictures of, then you only image the ground and skip the clouds. That way, you’re not storing, processing, and downloading all this imagery researchers really can’t use,” said Ben Smith, associate at the Earth Science Technology Office, during a media briefing.

NASA reports that up to 67 percent of Earth-observation scenes are obscured by cloud cover globally. The test demonstrated effective image filtering, with the system skipping clouded scenes entirely to conserve power and bandwidth.

Future upgrades will allow satellites to track wildfires, storms, and other transient phenomena, as well as coordinate across a fleet using inter-satellite communication under the Federated Autonomous MEasurement project.

JPL’s Pasadena-based team continues to advance autonomous space technologies, building on its prior work with missions like Earth-Observing-1 and Rosetta. Dynamic Targeting’s success marks a major step toward smarter, more agile scientific instruments.