Tracking of Maneuvering Targets with Poisson Labeled multi-Bernoulli Tracker

In this paper, the Poisson labeled multi-Bernoulli (PLMB) tracker is exploited to solve the problem of maneuvering target tracking. For avoiding tracking divergence resulting from a single motion model, multiple models (MM) strategy is applied to the PLMB tracker. Specifically, each state of target is coupled with all possible motion models and the model transition is considered throughout the tracking process. Simulation results demonstrate that the proposed MM-PLMB tracker strikes a favorable balance between limited computational resources and high tracking accuracy when compared to state-of-the-art approaches.