Track-to- Track Fusion for Elliptical Extended Targets Parameterized with Orientation and Semi-Axes Lengths

Track-to-track fusion considers the problem of fusing multiple tracks (e.g., from different sensor nodes) of the same target object. In case the spatial extent of the target object is estimated, unique challenges for the track-to-track fusion method arise, e.g., there may be ambiguities in the parameterization. In this work, we present an approach for distributed track-to-track fusion in case of elliptical extend targets, where correlations from common prior and process noise are explicitly incorporated. The approach is based on the previously introduced Random Ellipse Density (RED) framework, which deals with ambiguous ellipse representations and utilizes a Minimum Mean Gaussian Wasserstein (MMGW) estimator that is optimal with respect to the Gaussian Wasserstein (GW) distance. We provide simulated experiments in which measurements from different sensors are tracked by elliptic extended object trackers, fusing the trackers' estimates to improve the estimation.