Integrated multiple-hypothesis nonlinear tracking

Technologies for adaptive data fusion, multiple-hypothesis fusion management, and nonlinear tracking have, over the years, matured along separate lines of research and development. In this paper, we introduce an extendable, multi-source acoustic tracking system architecture for undersea C4ISR applications, which integrates a multiple-hypothesis tracking (MHT) paradigm with a Monte Carlo particle tracking paradigm. The former approach models target state probability densities with a Gaussian sum Iterated Extended Kalman Filter (IEKF), whereas the latter method employs non-linear probability state distributions. The prototype system, designated MultiStar, is being designed to apply the tracking methodology best suited to the estimated state probability of the targets being tracked. Preliminary analyses indicate that this integrated data fusion architecture will produce significantly better tracking estimates than either of the methods implemented separately.