Arbitrary Clutter PHD Filter and Its Implementation

Abstract

This paper is focus on multiple target tracking with arbitrary clutter process by using the probability hypothesis density (PHD) filter. Based on the Random Finite Set (RFS) framework, the clutter process in the classical PHD filter is modeled as Poisson RFS, which is only reasonable for some scenarios in reality. The PHD filter suitable for arbitrary clutter process has not yet been implemented. In this paper, an arbitrary clutter PHD filter, as well as its particle filter implementation, are derived by using the probability generating functional (PGFL) based on the RFS framework. Simulation results show that the proposed arbitrary clutter PHD filter can be used for multiple target tracking with arbitrary clutter process.