DTACSS: Decision-Theoretic Automation for Coordinated Space Surveillance

While the United States is increasingly reliant on space-based technologies, our satellite assets are increasingly vulnerable to attacks and collisions with debris or other satellites. Current space surveillance systems are stovepiped and unresponsive; when possible collisions ("orbital conjunctions") are detected in advance, additional tracking data can only be acquired through complex human interactions and requests among different organizations. Key challenges include managing uncertainty and utility with scarce resources, enforcing real-time response to near-term threats, and coordinating distributed teams of humans and autonomous systems. SIFT proposes to design DTACSS to meet the challenges of dynamic uncertainty management and active space surveillance control. In the DTACSS concept, intelligent agents will coordinate the activities of each element of the future space surveillance and response network (i.e., an evolution of the current Air Force Satellite Control Network). DTACSS agents will operate in both mixed-initiative and fully-autonomous modes, reasoning about threat probabilities and possible outcomes, collaborating with human operators to define appropriate mission priorities and a-priori response plans, and autonomously executing responses to long-term and real-time threats. BENEFIT: DTACSS will dramatically improve the effectiveness and survivability of military satellites by providing the following benefits: dynamic tasking of surveillance assets to explicitly manage uncertainty and risk, real-time response to threats, and coordinated command and control. There are a wide variety of potential military and civilian applications of the DTACSS concept, helping teams of humans and autonomous systems coordinate their functions in the face of uncertainty and limited resources.