Ground Station and Software Defined Satellites (SDS) Data Fusion to Improve Space Situational Awareness (SSA) Using Federated Learning (FL)

Prescient Edge proposes the use of a Decentralized Machine Learning model that incorporates real-time updates (called Federated Learning) to improve the current models that predict orbital paths of satellites, for improved Space Situational Awareness (SSA). Federated learning trains AI at the edge of a distributed network without the need for centralizing the data. In the space domain, this means giving satellites advanced AI capabilities without the need of a constant connection with ground stations.  Additionally, federated learning AI only communicates heuristics, thus keeping the underlying data secure from peer competitors. Prescient Edge’s novel federated learning AI is currently deployed in the healthcare industry. Prescient Edge is one of six federated learning AI companies based out of the US, and one of nine federated learning companies in the world.  With the exponential growth in both commercial and DoD space, the need for AI powered deconfliction and collision will address the dual-use market. Compared to traditional AI, Federated Learning is a great way to deploy AI in space because it is (1) more secure, (2) fewer processing requirements, and (3) requires less connectivity to a central node.