An Efficient Labeled/Unlabeled Random Finite Set Algorithm for Multiobject Tracking

In this article, we propose an efficient random finite set (RFS)-based algorithm for multiobject tracking, in which the object states are modeled by a combination of labeled multi-Bernoulli (LMB) RFS and Poisson RFS. The less computationally demanding part is used to track potential objects whose existence unlikely. Only if quantity characterizing plausibility above threshold, new Bernoulli component created, tracked more accurate but LMB algorithm. Conversely, transferred back corresponding probability falls below another threshold. Contrary existing hybrid algorithms based on RFSs, proposed method facilitates continuity implements complexity-reducing features. Simulation results demonstrate large complexity reduction relative other RFS-based with comparable performance.