A fuse-before-track approach to target state estimation using passive acoustic sensors

This paper describes a Bayesian fuse-before-track approach to detecting and tracking moving targets by fusing data from multiple passive acoustic arrays. We describe a surveillance application, where a collection of fixed location arrays are charged with monitoring a spatial region. Each array provides information only about target bearing relative to the array. The conventional approach is to track bearing at each array and then fuse the tracks to estimate XY position. Instead, in fuse-before-track, we fuse measurements before creating tracks. The fusion is done using a nonlinear filter, where non-thresholded measurements corrupted by non-Gaussian noise which are related non-linearly to the desired target state are combined from all arrays in one tracker. We illustrate the algorithms' efficacy on real, collected at-sea data.