Star convex irregular shape multi-extended target PMBM filter

This work suggests a star convex irregular shape multi-extended target Poisson multi-Bernoulli mixture (PMBM) filtering algorithm to address the challenge of monitoring multiple extended targets with irregular shapes in complicated and uncertain situations. First, the Poisson point process (PPP) and multi-Bernoulli mixture (MBM) are used to model the unknown and existing target sets, effectively representing potential target information while establishing an efficient multi-target density recursive form. The measurement source distribution of any star convex extended target is accurately modeled by the random hypersurface model, and the best nonlinear filter solves the highly nonlinear pseudo measurement equation. On the basis of deriving the extended target Poisson multi-Bernoulli mixture (ETPMBM) filter, the Gaussian mixture star convex irregular shape multi-extended target PMBM filter algorithm is derived and proposed in detail. In order to effectively and recursively estimate the probability density of numerous extended targets with multiple feature information, including irregular forms, this approach can create more compact multi-Bernoulli global hypotheses. Finally, the effectiveness of the algorithm proposed in this paper is verified through simulation experiments of multiple extended target tracking and multiple group target tracking.