Bayesian Fusion of Heterogeneous Data for Unanticipated Event Detection

Metron, Inc. and Portland State University propose to design, develop, test and demonstrate algorithms that fuse heterogeneous sensor data at an information theoretic level and process these data as detection and tracking algorithms that can identify when new events or targets emerge that are not currently modeled. These algorithms have the potential to provide dramatic improvements in the capability to detect and track in multi-sensor environments when unknown events or targets enter the scene.