DATA FUSION & NEURAL NETWORKS, LLC — Department of Defense SBIR Phase I: AF161-059

Description

DF&NN and AGI will leverage the available multi-INT data from ComSpOC and AFRL to understand patterns in currently evolving situations in space (i.e., entity/event relationship information) and compare them to available truth to provide threat indications and warnings (I&W) knowledge left of significant events. SSA community tools being developed now are working on the first step to provide information to the warfighter from multi-INT data sources for the characterization of space objects and the identification of specific events of interest. Our intelligent system will learn patterns of life (e.g., correlations in normal entity/event relationships) to detect threat-caused abnormal relationship patterns. Our proposed approach is to divide & conquer this entity/event relationship learning, detection, and characterization problem into four types. We will apply the appropriate problem-to-solution mappings to each and integrate these into the Event Relationship Recognition for SSA (ERRS) prototype.ComSpOC will provide information from a network of diverse phenomenology global sensors. The ERRS will provide entity/event and relationship recognition, relationship learning, abnormal relationship detection, mission impact assessment, and CACM. The ERRS prototype will address selected portions of these based upon the data. We will extend our Performance Assessment and Process Management (PAPM) tools to evaluate experiments on ERRS.