Concepts of Resident Objects and Situations in Space (CROSS)

The XAnalytix STTR Phase 2 proposal termed Concepts of Resident Objects and Situations in Space (CROSS) will assist commanders and decision-makers through various analysis and modeling techniques that automate the evaluation of alternative objects or situations of interest. The proposed program applies the Conceptual Spaces framework to the problem of RSO identification and space event characterization. Relevant domains, dimensions, and properties will be modeled through this paradigm. The geometric representations established through this process allow space object classifications to be produced from observations captured and directed to a Conceptual Spaces fusion engine built upon well-defined mathematical programming models. To alleviate the burden of manually constructing Conceptual Space models, explainable machine learning techniques will be applied to help automate this process. The system designed under CROSS will then be expanded to accommodate the construction and analysis of Situational Spaces, an extension of Conceptual Spaces that includes understanding property changes over time, relationships between objects, and detection of potential situations. The Situational Spaces concept is key for space event characterization, pushing the Conceptual Spaces state-of-the-art beyond object classification into higher-order understanding. For this second phase, the concentration will be on advancing the following areas: • Implementation of the CROSS Architecture designed in Phase I. • Expansion to Situational Spaces to provide situational understanding and plausible futures. • Development of mathematical model and corresponding scalable algorithms. • Finalizing a User Interface for both Conceptual and Situational Spaces. • Providing an efficient database structure that will allow for growth and usability. • Demonstration of the overall system with simulated and synthetic realistic scenarios.