A small spacecraft earlier this year successfully demonstrated edge autonomy — or the capability of a satellite to analyze data and make decisions independently without waiting for instructions from ground control.
The demonstration was conducted by Arlington, Virginia-based startups NOVI and Sedaro. NOVI, which specializes in edge computing systems that leverage AI and machine learning algorithms for satellite intelligence, built the spacecraft for an undisclosed U.S. government customer. The satellite launched to low Earth orbit on SpaceX’s Transporter-12 rideshare mission Jan. 14, with most of the demonstration activities completed in February.
To perform autonomous operations, the spacecraft’s onboard computer uplinked Sedaro’s mission and system simulation software, which functioned as a digital twin — a continuously updated virtual representation of the actual satellite in orbit.
In this demonstration, the Sedaro software enabled the satellite to be “self aware,” or to autonomously execute tasks based on its environment and status, the companies said.
Sedaro developed this technology under a SpaceWERX Small Business Innovation Research grant. SpaceWERX, the space-focused division of the Department of the Air Force’s AFWERX organization, is interested in adapting commercial edge autonomy technology for military satellite applications.
The technology allows satellites to independently adjust orbits, manage power consumption, or optimize communication links based on sensor data. In a military context, self-aware satellites could detect if they are being tracked or targeted, and respond accordingly.
Industry implications
Autonomous satellites enabled by AI and edge computing are seen as a cornerstone technology for the automation of space operations.
Sebastian Welsh, Sedaro’s co-founder and CTO, said satellite edge autonomy is “essential for sustained commercial success and strategic advantage in space.”
“The proliferation of constellations in low Earth orbit has driven increasingly automated approaches to satellite mission ops from the ground,” Welsh said April 3 in a news release. “A highly congested, contested, and dynamic space operating environment demands that we take the next step and bring this automation to the edge.”
Sedaro CEO Robbie Robertson added that the demonstration “validates an approach to autonomy that can deliver value in the very near future onboard government and commercial space assets.”
He said the software supporting the demonstration, called “Sedaro Autonomy Framework for the Edge,” will be released as open-source. This decision aims to “accelerate adoption and disrupt the extremely slow and expensive process of innovation that is the norm in space and defense.”
Amit Mehra, NOVI co-founder, said the company’s goal is to create an AI processing infrastructure in space “through a combination of low-cost, multi-sensor satellites and advanced edge processors.” This infrastructure could be used to train AI algorithms for commercial and defense applications, he said, and would provide third-party developers with access to space-based sensors and processing to develop, upload and operate their own algorithms.
“The Sedaro work was one such end-to-end demonstration of ground testing of a third-party algorithm, followed by seamless upload and in-space operation to obtain on-orbit data,” Mehra said.
