Classification Approach for Assessing Spacecraft Threats (CAAST)

Battlespace awareness within the space domain is a critical foundation for planning appropriate courses of action, responding to threats, protecting vulnerable assets, and preparing contingency plans. The ability to maintain flexible deterrent options in various situations relies upon maintaining an accurate picture of (1) what is happening now, and (2) what could happen in the future. Because the set of future possibilities is infinite, it is important to broadly characterize these possibilities and to best understand those which require a response or present the greatest threats to our assets. For this Phase I STTR, we propose to develop and test a Classification Approach for Assessing Spacecraft Threats (CAAST) that accounts for non-gaussian spacecraft state uncertainties and provides a probabilistic assessment of adversary co-orbital COAs as kinetic threat possible or kinetic threat not possible. We will develop a Support Vector Machine (SVM) Model with Gaussian Mixture uncertainties algorithm that accounts for adversary/asset state uncertainty, quantifies confidence in the threat assessment, and drastically reduces computation time compared with existing reachability solutions. The significant reduction in computation time will enable threat characterization for many adversary/asset pairs at once. Threat characterization on this scale is not currently possible due to computational intractability of existing tools. Our Phase I solution will focus on impulsive maneuvers at the start of the adversary trajectory, will focus on 2-body dynamics with use cases originating beyond GEO and terminating in GEO, and will classify threats as kinetic threat possible vs. kinetic threat not possible. Our Phase II solution will expand on the Phase I proof-of-concept to encompass all adversary thrust profiles, three body dynamics, all orbital regimes, and threat classification between kinetic threat, non-kinetic threat, and non-threatening. Our combined Trusted Space and CUBRC team has significant experience developing tools for spacecraft threat assessment, reachability analysis, and hypothesis management. Our teams also have experience working in operational Aerospace environments and bring a deep understanding of operator needs to this proposed work.