Feature-Enhanced Multiple Hypothesis Nonlinear Tracking

ORINCON and Metron propose to develop an integrated multiple hypothesis nonlinear tracking capability that crosses multiple feature (or attribute) domains while properly representing dependencies between these feature domains and the kinematic domain. Inprevious research, ORINCON and Metron successfully integrated a Bayesian particle filter tracker (Nodestar) and a multiple hypothesis, Gaussian sum, extended Kalman filter tracker (ALFP) into the MultiStar tracker, and demonstrated resulting improvementson passive sonar kinematic data. We will extend our MultiStar capability to non-kinematic attribute domains, and will define the appropriate architecture for properly representing interactions between the kinematic and non-kinematic domains. We willdemonstrate dependent attribute interactions on specifically targeted features. Initially, we will focus on passive narrowband frequency attributes provided by the state-of-the-art Acoustic Rapid COTS Insertion (ARCI) and ARCI(IUSS) contact followers andby the novel Spectral Histogram Probabilistic Multi-hypothesis Tracker (Spectral H-PMHT) contact follower research currently in progress. We will also address attributes provided by visual and radar/ESM sensors. Fusion that properly exploits theincreased observability across these attribute domains holds promise for immediate and substantial improvement in fused scene representation and management. We expect that this effort will result in practical implementation of a superior data fusion andtracking system that will be the standard against which other such systems will be compared. It will provide the increased observability, speed, and accuracy required to significantly improve the Fleet's capability to solve difficult problems in C4ISR,including close encounter, log likelihood tracking problems, and a broad range of track-before-detect problems. While the immediate application is to provide the highest quality data fusion and tracking system to the Navy for tactical use, the techniquesdeveloped will be applicable to a wide range of data fusion and tracking problems both in the military and civilian markets, including Integrated ASW, the Advanced Rapid COTS Insertion (ARCI) program, the Advanced Processor Build-Tactical (APB-T), and theARCI(I) program.