Advanced Feature Aided Track and CDI Fusion Processing of Data from

Successful missile defense operations require an accurate real-time single integrated battlespace picture, while minimizing data throughput and bandwidth requirements. Current multi-sensor tracking approaches that treat kinematic and feature fusion as separate processes have high rates of correlation and target typing errors. Our integrated fusion process integrates optical and radar data, thereby reducing battlespace picture uncertainty by jointly considering kinematic, radar features, and thermal imaging data. Networked sensors working jointly substantially increase lethal object identification capability by sharing independent 2-D or 3-D kinematic data and feature data such as RCS, IR signatures, and length profiles. This approach improves system performance by increasing the ability to resolve high speed closely spaced objects, track maneuvering objects (boosting missiles and maneuvering/unstable reentry vehicles), and identify and destroy lethal objects within dense target clusters containing decoys and debris. Intelligent Systems Research will develop an integrated and distributed target type fusion process within a distributed component system architecture that receives statistically independent data from distributed network sensors and simultaneously processes this data. Our approach operates on sensor reports containing features, tracklets, and object type likelihoods. We minimize network throughput by adaptively transmitting data with a high value-of-information according to the needs of network participants.