Fusion in a Cloud Navy SBIR FY2013.2 Sol No.: Navy SBIR FY2013.2 Topic No.: N132-135 Topic Title: Fusion in a Cloud Proposal No.: N132-135-0079 Firm: Black River Systems Company, Inc. 162 Genesee Street Utica, New York 13502 Contact: Peter Shea Phone: (315) 732-7385 Web Site: www.brsc.com Abstract: One of the major benefits of tactical cloud computing is improved net-centric capabilities and operations with the objective of information dominance. The reassign-able pools of computing resources and efficient information across operational boundaries can produce a more accurate and up to date common operational picture for warfighters. Instead of individual platforms pulling data from sensors to fuse and analyze locally for battle space and situational awareness, warfighters will be able to increase the accuracy of local scenes through remote data processing and sharing. Black River Systems Company will leverage its existing Distributed Fusion Manager (DFM) for deployment of distributed Level 1 and 2 fusion algorithms to the cloud. The DFM currently performs distributed Level 1 fusion by synchronizing track identities and performing track to track fusion over a network of tracking platforms. The DFM will be enhanced to provide an infrastructure for Level 2 fusion that will combine both probabilistic and non-probabilistic learning and inferencing models across geographically separated nodes. At the conclusion of Phase I, Level 1 fusion algorithms will be demonstrated using simulated Surface-Moving-Target-Indicator (SMTI) and radar tracks; Level 2 fusion algorithms will be demonstrated using relevance vector machine, support vector machine, and Bayesian network models. Benefits: The Data Fusion Manager (DFM) will provide the capability to perform real-time Level 1 and Level 2 distributed fusion in a tactical cloud for a wide variety of data sources and scenarios, providing increased situation awareness. The DFM Level 1 and proposed Level 2 fusion algorithms operate on the output of the local tracking, inference, and learning algorithms. They will support any probabilistic or non-probabilistic classification or inference technique-providing a versatile fusion capability for the tactical cloud that can be used on a wide variety of tracking, classification, and inference techniques. Return