Skip to Main Content
This paper describes a Bayesian approach to detecting and tracking multiple moving targets using acoustic data from multiple passive arrays. Traditional undersea acoustic systems develop tracks at the single array level, requiring track association between nodes with nonlinear projections from measurement space to target space. In contrast, our nonlinear filtering approach fuses data at the measurement level and operates directly in the target state space. As such, this approach directly addresses both the nonlinear sensor to target state coupling as well as the ambiguities caused by bearings-only nature of the passive regime. In particular, our method better addresses these challenges by combining high-fidelity physics-based sensor statistical modeling, an innovative nonlinear Bayesian filter, and a unique method of handing the computational implementation.