Augmented LRFS-based filter: Holistic tracking of group objects

Abstract

Aiming at the problem of accurate tracking of group objects, where multiple closely spaced objects within a group pose a coordinated motion, this paper develops a new type of labeled random finite set (LRFS), i.e., augmented LRFS, which inherently integrates group information such as the group geometry center and the group index into the definition of LRFS. Specifically, for each element in an augmented LRFS, the kinetic states, the track labels, and the corresponding group information of its represented object are incorporated. Then, by means of the proposed augmented LRFS-based filter, i.e., the labeled multi-Bernoulli filter with the proposed augmented LRFS, the group structure is iteratively propagated and updated during the tracking process, which achieves the holistic estimation of the kinetic states, track labels, and the corresponding group information of multiple group objects, and further improves the tracking performance. Finally, simulation experiments are conducted to verify the effectiveness of the proposed augmented LRFS-based filter.