SBIR-STTR Award

In this work we propose an integrated Finite Set Statistical (FISST) and hierarchical reasoning filtering solution to enable rapid catalog processing and object characterization for threat indication and warning. The solution relies on defining a hybrid discrete-continuous relational state between an asset and all potentially threatening space objects, and then applied to all assets. Given existing representative catalog, and raw sensor and other soft data, FISST allows for the extraction of the most amount of information contained in the data. This allows for a more accurate catalog processor leading to more rapid object assessment and characterization. Such accuracy and rapidity aids in faster and more accurate threat indications and warning. The proposed solution is probabilistically rigorous, scalable and can be implemented in real-time.;

Benefit:
This Phase I effort will be foundational to future work efforts by establishing a working prototype that exercises a variety of new probabilistically rigorous and scalable hybrid filtering techniques for threat indications and warning.It will begin with the divergence from the standard paradigm, where catalog processing is completely divorced from object characterization and vice-versa, into a functioning, scalable and rapid tool for threat indications and warning.This paradigm change is essential to evolve to higher levels of rapid object characterization and potentially orders of magnitude improvement in threat indication and warning.The prototype developed in this effort will demonstrate how to effectively implement the innovative FISST methodology in an ARCADE SOA environment.The inherent capabilities of FISST and hierarchical reasoning, in regards to probabilistic and confidence calculations, may then be exercised within the ARCADE and intelligence community counterparts to evaluate how they can influence operational commanders perspective on SSA.Finally, this work would demonstrate a viable set of SOA-available inference algorithms that are sensor and data type agnostic, real-time responsive, and scalable enough to support the massive new data rates of systems (e.g. Lockheed Space Fence) that the JMS will acquire in next few years.This project would set the ground work for a series of technologies that could significantly increase the capabilities of the JSpOC/JMS to help achieve their missions in SSA, Force Protection, and Combat Identification.

To further enhance our ability to assess orbital events in the context of highly complex and ever evolving situations, provide a fuller understanding of the space operational picture, support indications and warnings, and to provide tactical protection of U.S. space assets, ADS has assembled a world class team of experts in the fields of Finite Set Statistics (FISST) and probabilistic reasoning for resident space object (RSO) recognition and identification. JMS requires a scalable cataloging architecture that accommodates high volumes of real-time raw measurement data processing and enable the JSpOC to appropriately balance their surveillance resources between catalog maintenance and tactical monitoring. Space event analysis is typically treated as the detection, tracking, identification, and characterization of individual RSOs where we seek to assess and analyze their interactions and patterns of behavior to ascertain potential threats. The ADS team presents an innovative information theoretic filtering methodology that allows the detection, correlation and threat characterization of pair-wise asset-RSO space threats. In this work, we apply powerful and intuitive hierarchy constructs to jointly reason between threat considerations, object characteristics, and mission specific behaviors. To this end, ADS proposes the incremental design and development of a scalable, automated rapid cataloging and characterization system called Semnai.