SBIR-STTR Award

The development and integration of three unique and innovative algorithm prototypes into a ?Fused Track and Characterization Schema are proposed. This Schema will encompass the determination of signatures and characteristics of objects that can be identified by RF and EO/IR Sensors in order to enable multi-sensor data fusion and correlation. The first algorithm, the 3D Pose Estimation Algorithm, permits track and correlation of multiple objects using multisensory EO/IR data measurements. The second algorithm, ?Wager? correlates multiple RF and sensor measurements with the EOIR ?Pose? solution to generate system level tracks. ?Wager? is based on automated Best Estimate of Truth methods. The third algorithm, ?Fused TASM? uses fused Target Attributes Surface Manifold Classifiers to characterize and associate targets across multiple sensors. TASM Classifiers, inherently by design, embed physical relationships between multiple sensor phenomenologies. The TASM Classifier construction is based on advanced Computer Graphics NURBS methods.

Keywords:
Multi-Phenomenology Sensor Data Fusion; Track Correlation And Association; Surface Manifold Classifiers; Discrimination; 3d Pose Estimation.

Under the Phase I effort, a set of algorithms were developed/enhanced that, when integrated into a fused track and characterization schema, are capable of realizing the full potential performance afforded by the battle manager having multiple sensors. dBWager is a multi-sensor measurement correlation algorithm that provides highly accurate state vector estimates and provides the correlation of electro-optic/infrared (EO/IR) and radiofrequency (RF) measurements to facilitate a true fused classification algorithm. dBTASM provides rapid RF object classification. 3D-Pose provides three dimensional (3D) position recovery from passive two dimensional (2D) imagery. Under the Phase II effort these prototype algorithms will be matured and evaluated in the Collaborative Test and Evaluation Center (CTEC). Additionally, new classification EO/IR algorithms will be developed, based on feature surface manifolds identified in the Phase I effort, that leverage the early RF classification that dBTASM provides. Approved for Public Release 14-MDA-7739 (18 March 14).

Keywords:
Multi-Phenomenology Sensor Data Fusion, Track Correlation and Association, Surface Manifold Classifiers, Discrimination, 3D Pose Estimation