Airborne Graph Analytics Applications for Multi-sensor Fusion and Integration

ABSTRACT: Black River Systems will develop a graph based software prototype that will demonstrate the following capabilities: (1) Capturing, maintaining, and updating a graph based representation of entity association ambiguities of an onboard integration and fusion capability for kinematic and identification sensor information; (2) An automated real time graph solver that produces the most likely set of stitched track trajectories that are consistent with all of the kinematic and identification collected up to the current time, as well as updating the entire history of tracks as new informationboth kinematic and identificationis collected and dynamically added to the graph; (3) Provide the feasibility of solving the graph that will be evaluated through the use of Government tracking metrics; and (4) Demonstrate the feasibility of performing post mission forensic analysis by combining other kinematic and identification sources based on new evidence obtained after the mission has ended.; BENEFIT: Our graphical based representation of ambiguous multi-INT associations provides a flexible graph structure that is solved in real time and preserves all target associations until future information can be used for target resolution.The real-time solver produces the best tracks on demand, but uses the un-pruned graph to determine better fused target trajectories over time as new information is added.Based on design feedback from real world operators analyzing data for network discovery, our approach builds operator confidence in in the provided solutions.Finally, post mission analysis is supported through the capability of adding other mission collections to the inflight generated graph.