By Theresa Hitchens,
Published by Breaking Defense, 10 February 2022
On-board, or “edge,” processing is key to SDA’s plans to develop a satellite network that can gather data from multiple sensors on the ground, at sea and in space to provide operators with precise, near-real-time targeting information.
WASHINGTON: A foundational effort to test software for the Space Development Agency’s planned National Defense Space Architecture satellites successfully has wrapped up an initial demonstration of an on-board, automated data fusion capability, agency and industry sources said.
On-board, or “edge,” processing — rather than being done at ground stations, which slows down information flow — is key to SDA’s plans to develop a satellite network that can gather data from multiple sensors on the ground, at sea and in space to provide operators with precise, near-real-time targeting information.
Demonstrating that capability, which will require artificial intelligence/machine learning software, is the job of the Prototype On-orbit Experimental Test Bed (POET), launched aboard a Loft Orbital satellite this past June.
POET is testing critical software and hardware elements for SDA’s initial set of 28 Tranche 0 (T0) satellites, as well as later iterations, to be based in Low Earth Orbit (LEO). Tranche 0 includes early-design birds for SDA’s Tracking Layer which will detect and track ballistic and hypersonic missiles, and its backbone communications/relay Transport Layer constellation. Those satellites are to be launched beginning in late 2022.
“POET will demonstrate integration of a third-party multiple intelligence (multi-INT) data fusion software application in a LEO satellite modular and upgradeable mission software suite running in an edge-processor,” a June SDA press release announcing the launch explained.
The recent test, announced Feb. 8, used Scientific Systems Co. Inc.’s (SSCI) Collaborative Mission Autonomy (CMA) software that was originally tailored for DARPA’s Blackjack Pit Boss project. Blackjack is aimed at developing onboard battle management, command, control, and communications (BMC3) capabilities for large LEO constellations. SDA took over the software development last spring, and integrated it into POET.
POET also is testing a baseline avionics software processor, the CFC-400 built by Innoflight, to be carried on four Tranche 0 Tracking Layer satellites being built by L3Harris under a 2020 contract. (SpaceX is also building four.)
“The most important reason for the POET experiment was to test the Innoflight CFC-400 processor in the on-orbit environment to mitigate risk for T0 spacecraft flying related hardware,” an SDA official explained. “POET is a prototype testbed, as its name implies. Results from POET provide data that can be used to inform future decisions.”
During the initial demo, SSCI’s CMA software, running on the CFC-400 processor, integrated a cloud detection application that uses imaging data collected by an electro-optical camera built by BlackSky, explained a Feb. 8 SSCI press release.
SDA’s planned Tracking Layer constellation in Low Earth Orbit (LEO) for missile warning will need to be able to automatically detect clouds and “clean” them from tracking images provided to operators.
“The initial tests with POET focused on on-orbit cloud segmentation algorithms because those were readily available,” the SDA official said.
Owen Brown, SSCI’s vice president of solutions development, told Breaking Defense, that this first experiment was “focused on the data exploitation elements.”
Brown noted that the POET experimental satellite will remain on orbit and operational for five years, meaning that there are many possible experiments that could be done in the future.
The SDA official said “additional experiments/algorithms are being investigated to further test the POET processors,” but “no decisions have been made at this time.”
See: Original Article