Data Association for Cardinalized Probability Hypothesis Density Filter

Abstract

The main drawback of the cardinalized probability hypothesis density (CPHD) filter is that it can't identify the trajectories of different targets. A data association method, the CPHD filter combined with joint probabilistic data association (JPDA), is presented to track multiple targets in dense clutter. The CPHD filter is used as a pre-filter to remove unlikely measurements before inputting the remaining data to JPDAF for implementing data association. Track initiation and termination logic are employed to confirm the tracks and consequently ensure the implementation of JPDAF. Simulation results show that this approach works well in dense cluttered environments.