Poisson multi-Bernoulli filter for extended object tracking

Preprint, 2016

This paper presents a track-oriented Poisson multi-Bernoulli (PMB) filter for extended objects. The PMB filter is based on the Poisson multi-Bernoulli mixture (PMBM) conjugate prior and approximates the posterior PMBM by merging tracks across data association hypotheses. A method to create new tracks in a reasonable manner is proposed which uses a combination of pre-clustering, recycling and equivalence class among PMBM distributions.To approximate the marginal distributions of different tracks we use Gibbs sampling. The performance of the PMB is compared to the PMBM using a simulated scenario.