The 511 Space Innovation Team has taken their ISS-COMLSP Artificial Intelligence (AI) Agent to new “heights” with an in-flight demo onboard the International Space Station (ISS).
The ISS–Communications Onboard Machine Learning Signal Predictor, or ISS-COMLSP, is a slightly modified version of their ground AI Agent, the IAM-MLSP, currently being developed for ISS operations in Houston’s Mission Control Center.
The IAM-MLSP, or ISS Antenna Manager-Machine Learning Signal Predictor, predicts upcoming loss of signal (LOS) events and feeds that information into the IAM console tool used by flight controllers to manage TDRS satellite handovers and monitor for LOS situations.
“Our AI LOS prediction technology removes the ambiguity and the subjective human nature of LOS calls” says Dr. Tim Giblin, 511 Space Innovation Founder and CEO. “Such a higher dimensional problem poses interpretive problems for the human mind to swiftly solve in real-time, but can be solved with artificial intelligence.”
Dr. Steve Novotny, 511’s Director of AI and ML, incorporated cutting edge AI techniques based on transformer networks to be able to handle higher dimensional time series data imposed by this problem.
Through a NASA SBIR Award, 511 Space Innovation is working to finalize a prototype IAM-MLSP to be tested in the ISS Mission Support Development Environment at NASA’s Johnson Space Center.
“Ultimately, our team wants to embed this LOS predictor technology into avionics systems onboard spacecraft. Not only will this allow the onboard crew to have prior knowledge of forthcoming LOS instances, but the output can also feed into the spacecraft’s autonomous systems” says IAM-MLSP principal investigator and lead developer, Dr. Brice Orange.
Lat week 511 Space Innovation took their first step toward making that a reality.
In collaboration with Hewlett Packard Enterprise’s Spaceborne Computer-2 (SBC-2), the 511 Team’s ISS-COMLSP was uplinked to ISS and successfully executed on Thursday, April 3rd 2025.
Although no crew interface to the SBC-2 exists, HPE’s goal with SBC-2 is to demonstrate and analyze the requirements for edge computing in space systems by test running various AI Agents. 511’s ISS-COMLSP is one of only a handful of AI Agents to run onboard the International Space Station.
“This is an amazing first step for the 511 Team. We are now getting a better handle on the computational demands for an embedded intelligent system” remarked Dr. Orange.
As space exploration moves beyond Low Earth Orbit (LEO) in the coming decade, space communication networks will become increasingly complex, and likely riddled with unsuspecting LOS events. 511’s cutting edge LOS predictor technology will be at the forefront to minimize, or perhaps mitigate altogether, the negative impacts of LOS on in-flight operations.
For more information, or to discuss how an LOS Predictor AI Agent might be integrated into your spacecraft system, please contact Dr. Tim Giblin Timothy.Giblin@511SpaceInnovation.com
6 April 2025
The 511 Space Innovation Team has taken their ISS-COMLSP Artificial Intelligence (AI) Agent to new “heights” with an in-flight demo onboard the International Space Station (ISS).
The ISS–Communications Onboard Machine Learning Signal Predictor, or ISS-COMLSP, is a slightly modified version of their ground AI Agent, the IAM-MLSP, currently being developed for ISS operations in Houston’s Mission Control Center.
The IAM-MLSP, or ISS Antenna Manager-Machine Learning Signal Predictor, predicts upcoming loss of signal (LOS) events and feeds that information into the IAM console tool used by flight controllers to manage TDRS satellite handovers and monitor for LOS situations.
“Our AI LOS prediction technology removes the ambiguity and the subjective human nature of LOS calls” says Dr. Tim Giblin, 511 Space Innovation Founder and CEO. “Such a higher dimensional problem poses interpretive problems for the human mind to swiftly solve in real-time, but can be solved with artificial intelligence.”
Dr. Steve Novotny, 511’s Director of AI and ML, incorporated cutting edge AI techniques based on transformer networks to be able to handle higher dimensional time series data imposed by this problem.
Through a NASA SBIR Award, 511 Space Innovation is working to finalize a prototype IAM-MLSP to be tested in the ISS Mission Support Development Environment at NASA’s Johnson Space Center.
“Ultimately, our team wants to embed this LOS predictor technology into avionics systems onboard spacecraft. Not only will this allow the onboard crew to have prior knowledge of forthcoming LOS instances, but the output can also feed into the spacecraft’s autonomous systems” says IAM-MLSP principal investigator and lead developer, Dr. Brice Orange.
Lat week 511 Space Innovation took their first step toward making that a reality.
In collaboration with Hewlett Packard Enterprise’s Spaceborne Computer-2 (SBC-2), the 511 Team’s ISS-COMLSP was uplinked to ISS and successfully executed on Thursday, April 3rd 2025.
Although no crew interface to the SBC-2 exists, HPE’s goal with SBC-2 is to demonstrate and analyze the requirements for edge computing in space systems by test running various AI Agents. 511’s ISS-COMLSP is one of only a handful of AI Agents to run onboard the International Space Station.
“This is an amazing first step for the 511 Team. We are now getting a better handle on the computational demands for an embedded intelligent system” remarked Dr. Orange.
As space exploration moves beyond Low Earth Orbit (LEO) in the coming decade, space communication networks will become increasingly complex, and likely riddled with unsuspecting LOS events. 511’s cutting edge LOS predictor technology will be at the forefront to minimize, or perhaps mitigate altogether, the negative impacts of LOS on in-flight operations.
For more information, or to discuss how an LOS Predictor AI Agent might be integrated into your spacecraft system, please contact Dr. Tim Giblin Timothy.Giblin@511SpaceInnovation.com